Training in progress, step 1250, checkpoint
Browse files- checkpoint-1250/README.md +202 -0
- checkpoint-1250/adapter_config.json +34 -0
- checkpoint-1250/adapter_model.safetensors +3 -0
- checkpoint-1250/global_step1249/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt +3 -0
- checkpoint-1250/global_step1249/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt +3 -0
- checkpoint-1250/global_step1249/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt +3 -0
- checkpoint-1250/global_step1249/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt +3 -0
- checkpoint-1250/global_step1249/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt +3 -0
- checkpoint-1250/global_step1249/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt +3 -0
- checkpoint-1250/global_step1249/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt +3 -0
- checkpoint-1250/global_step1249/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt +3 -0
- checkpoint-1250/global_step1249/mp_rank_00_model_states.pt +3 -0
- checkpoint-1250/latest +1 -0
- checkpoint-1250/rng_state_0.pth +3 -0
- checkpoint-1250/rng_state_1.pth +3 -0
- checkpoint-1250/rng_state_2.pth +3 -0
- checkpoint-1250/rng_state_3.pth +3 -0
- checkpoint-1250/rng_state_4.pth +3 -0
- checkpoint-1250/rng_state_5.pth +3 -0
- checkpoint-1250/rng_state_6.pth +3 -0
- checkpoint-1250/rng_state_7.pth +3 -0
- checkpoint-1250/scheduler.pt +3 -0
- checkpoint-1250/special_tokens_map.json +30 -0
- checkpoint-1250/tokenizer.json +0 -0
- checkpoint-1250/tokenizer_config.json +133 -0
- checkpoint-1250/trainer_state.json +2308 -0
- checkpoint-1250/training_args.bin +3 -0
- checkpoint-1250/zero_to_fp32.py +674 -0
checkpoint-1250/README.md
ADDED
|
@@ -0,0 +1,202 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
base_model: microsoft/Phi-3-mini-4k-instruct
|
| 3 |
+
library_name: peft
|
| 4 |
+
---
|
| 5 |
+
|
| 6 |
+
# Model Card for Model ID
|
| 7 |
+
|
| 8 |
+
<!-- Provide a quick summary of what the model is/does. -->
|
| 9 |
+
|
| 10 |
+
|
| 11 |
+
|
| 12 |
+
## Model Details
|
| 13 |
+
|
| 14 |
+
### Model Description
|
| 15 |
+
|
| 16 |
+
<!-- Provide a longer summary of what this model is. -->
|
| 17 |
+
|
| 18 |
+
|
| 19 |
+
|
| 20 |
+
- **Developed by:** [More Information Needed]
|
| 21 |
+
- **Funded by [optional]:** [More Information Needed]
|
| 22 |
+
- **Shared by [optional]:** [More Information Needed]
|
| 23 |
+
- **Model type:** [More Information Needed]
|
| 24 |
+
- **Language(s) (NLP):** [More Information Needed]
|
| 25 |
+
- **License:** [More Information Needed]
|
| 26 |
+
- **Finetuned from model [optional]:** [More Information Needed]
|
| 27 |
+
|
| 28 |
+
### Model Sources [optional]
|
| 29 |
+
|
| 30 |
+
<!-- Provide the basic links for the model. -->
|
| 31 |
+
|
| 32 |
+
- **Repository:** [More Information Needed]
|
| 33 |
+
- **Paper [optional]:** [More Information Needed]
|
| 34 |
+
- **Demo [optional]:** [More Information Needed]
|
| 35 |
+
|
| 36 |
+
## Uses
|
| 37 |
+
|
| 38 |
+
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
|
| 39 |
+
|
| 40 |
+
### Direct Use
|
| 41 |
+
|
| 42 |
+
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
|
| 43 |
+
|
| 44 |
+
[More Information Needed]
|
| 45 |
+
|
| 46 |
+
### Downstream Use [optional]
|
| 47 |
+
|
| 48 |
+
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
|
| 49 |
+
|
| 50 |
+
[More Information Needed]
|
| 51 |
+
|
| 52 |
+
### Out-of-Scope Use
|
| 53 |
+
|
| 54 |
+
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
|
| 55 |
+
|
| 56 |
+
[More Information Needed]
|
| 57 |
+
|
| 58 |
+
## Bias, Risks, and Limitations
|
| 59 |
+
|
| 60 |
+
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
|
| 61 |
+
|
| 62 |
+
[More Information Needed]
|
| 63 |
+
|
| 64 |
+
### Recommendations
|
| 65 |
+
|
| 66 |
+
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
|
| 67 |
+
|
| 68 |
+
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
|
| 69 |
+
|
| 70 |
+
## How to Get Started with the Model
|
| 71 |
+
|
| 72 |
+
Use the code below to get started with the model.
|
| 73 |
+
|
| 74 |
+
[More Information Needed]
|
| 75 |
+
|
| 76 |
+
## Training Details
|
| 77 |
+
|
| 78 |
+
### Training Data
|
| 79 |
+
|
| 80 |
+
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
|
| 81 |
+
|
| 82 |
+
[More Information Needed]
|
| 83 |
+
|
| 84 |
+
### Training Procedure
|
| 85 |
+
|
| 86 |
+
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
|
| 87 |
+
|
| 88 |
+
#### Preprocessing [optional]
|
| 89 |
+
|
| 90 |
+
[More Information Needed]
|
| 91 |
+
|
| 92 |
+
|
| 93 |
+
#### Training Hyperparameters
|
| 94 |
+
|
| 95 |
+
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
|
| 96 |
+
|
| 97 |
+
#### Speeds, Sizes, Times [optional]
|
| 98 |
+
|
| 99 |
+
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
|
| 100 |
+
|
| 101 |
+
[More Information Needed]
|
| 102 |
+
|
| 103 |
+
## Evaluation
|
| 104 |
+
|
| 105 |
+
<!-- This section describes the evaluation protocols and provides the results. -->
|
| 106 |
+
|
| 107 |
+
### Testing Data, Factors & Metrics
|
| 108 |
+
|
| 109 |
+
#### Testing Data
|
| 110 |
+
|
| 111 |
+
<!-- This should link to a Dataset Card if possible. -->
|
| 112 |
+
|
| 113 |
+
[More Information Needed]
|
| 114 |
+
|
| 115 |
+
#### Factors
|
| 116 |
+
|
| 117 |
+
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
|
| 118 |
+
|
| 119 |
+
[More Information Needed]
|
| 120 |
+
|
| 121 |
+
#### Metrics
|
| 122 |
+
|
| 123 |
+
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
|
| 124 |
+
|
| 125 |
+
[More Information Needed]
|
| 126 |
+
|
| 127 |
+
### Results
|
| 128 |
+
|
| 129 |
+
[More Information Needed]
|
| 130 |
+
|
| 131 |
+
#### Summary
|
| 132 |
+
|
| 133 |
+
|
| 134 |
+
|
| 135 |
+
## Model Examination [optional]
|
| 136 |
+
|
| 137 |
+
<!-- Relevant interpretability work for the model goes here -->
|
| 138 |
+
|
| 139 |
+
[More Information Needed]
|
| 140 |
+
|
| 141 |
+
## Environmental Impact
|
| 142 |
+
|
| 143 |
+
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
|
| 144 |
+
|
| 145 |
+
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
|
| 146 |
+
|
| 147 |
+
- **Hardware Type:** [More Information Needed]
|
| 148 |
+
- **Hours used:** [More Information Needed]
|
| 149 |
+
- **Cloud Provider:** [More Information Needed]
|
| 150 |
+
- **Compute Region:** [More Information Needed]
|
| 151 |
+
- **Carbon Emitted:** [More Information Needed]
|
| 152 |
+
|
| 153 |
+
## Technical Specifications [optional]
|
| 154 |
+
|
| 155 |
+
### Model Architecture and Objective
|
| 156 |
+
|
| 157 |
+
[More Information Needed]
|
| 158 |
+
|
| 159 |
+
### Compute Infrastructure
|
| 160 |
+
|
| 161 |
+
[More Information Needed]
|
| 162 |
+
|
| 163 |
+
#### Hardware
|
| 164 |
+
|
| 165 |
+
[More Information Needed]
|
| 166 |
+
|
| 167 |
+
#### Software
|
| 168 |
+
|
| 169 |
+
[More Information Needed]
|
| 170 |
+
|
| 171 |
+
## Citation [optional]
|
| 172 |
+
|
| 173 |
+
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
|
| 174 |
+
|
| 175 |
+
**BibTeX:**
|
| 176 |
+
|
| 177 |
+
[More Information Needed]
|
| 178 |
+
|
| 179 |
+
**APA:**
|
| 180 |
+
|
| 181 |
+
[More Information Needed]
|
| 182 |
+
|
| 183 |
+
## Glossary [optional]
|
| 184 |
+
|
| 185 |
+
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
|
| 186 |
+
|
| 187 |
+
[More Information Needed]
|
| 188 |
+
|
| 189 |
+
## More Information [optional]
|
| 190 |
+
|
| 191 |
+
[More Information Needed]
|
| 192 |
+
|
| 193 |
+
## Model Card Authors [optional]
|
| 194 |
+
|
| 195 |
+
[More Information Needed]
|
| 196 |
+
|
| 197 |
+
## Model Card Contact
|
| 198 |
+
|
| 199 |
+
[More Information Needed]
|
| 200 |
+
### Framework versions
|
| 201 |
+
|
| 202 |
+
- PEFT 0.14.0
|
checkpoint-1250/adapter_config.json
ADDED
|
@@ -0,0 +1,34 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"alpha_pattern": {},
|
| 3 |
+
"auto_mapping": null,
|
| 4 |
+
"base_model_name_or_path": "microsoft/Phi-3-mini-4k-instruct",
|
| 5 |
+
"bias": "none",
|
| 6 |
+
"eva_config": null,
|
| 7 |
+
"exclude_modules": null,
|
| 8 |
+
"fan_in_fan_out": false,
|
| 9 |
+
"inference_mode": true,
|
| 10 |
+
"init_lora_weights": true,
|
| 11 |
+
"layer_replication": null,
|
| 12 |
+
"layers_pattern": null,
|
| 13 |
+
"layers_to_transform": null,
|
| 14 |
+
"loftq_config": {},
|
| 15 |
+
"lora_alpha": 16,
|
| 16 |
+
"lora_bias": false,
|
| 17 |
+
"lora_dropout": 0.0,
|
| 18 |
+
"megatron_config": null,
|
| 19 |
+
"megatron_core": "megatron.core",
|
| 20 |
+
"modules_to_save": null,
|
| 21 |
+
"peft_type": "LORA",
|
| 22 |
+
"r": 8,
|
| 23 |
+
"rank_pattern": {},
|
| 24 |
+
"revision": null,
|
| 25 |
+
"target_modules": [
|
| 26 |
+
"gate_up_proj",
|
| 27 |
+
"o_proj",
|
| 28 |
+
"qkv_proj",
|
| 29 |
+
"down_proj"
|
| 30 |
+
],
|
| 31 |
+
"task_type": "CAUSAL_LM",
|
| 32 |
+
"use_dora": false,
|
| 33 |
+
"use_rslora": false
|
| 34 |
+
}
|
checkpoint-1250/adapter_model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f43e012f1c21d754a50b38bef3b896c9bf9d6c4df42194652a2f5bd31711bcac
|
| 3 |
+
size 25200088
|
checkpoint-1250/global_step1249/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5362cb9d7f842b17d5275baf60140753351bb89ee4ae667557988eb0e959567e
|
| 3 |
+
size 18881328
|
checkpoint-1250/global_step1249/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:54504554c85f0311a58c1d23c42e7af254c38dfb447949f02ca01990613cb435
|
| 3 |
+
size 18881328
|
checkpoint-1250/global_step1249/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7b43f9dc9842bed261d6816fe3d9ef7bed7b6acb406087a9917abf87ec74c8a1
|
| 3 |
+
size 18881328
|
checkpoint-1250/global_step1249/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d05279d9adeaf91463a70e14823a9d59a485d5f484619a9dcd0a0b4dbdb6f7cb
|
| 3 |
+
size 18881392
|
checkpoint-1250/global_step1249/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:70d05101affbc5cd1c82cbbf0890c52cc9040bbe64c8dbb60986f08f30032e56
|
| 3 |
+
size 18881392
|
checkpoint-1250/global_step1249/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:743bddbee624936dd5534fcd2bcfd0447aaea997d4f8c2b6d83ccd8eb727131c
|
| 3 |
+
size 18881392
|
checkpoint-1250/global_step1249/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:85b489b85de118e7916fbe4c77004a9b9e2d3092ae01dd3a9b7cb8764c1196c8
|
| 3 |
+
size 18881392
|
checkpoint-1250/global_step1249/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f7217b92a2ce7e6adf02aec5c81aee6b307160c0b09d74468fa7669ca10375ef
|
| 3 |
+
size 18881392
|
checkpoint-1250/global_step1249/mp_rank_00_model_states.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b6eb63fc553d590b3f478cfe61f3aba2a1c0d71cc626f7e076eefbdf5a043c9f
|
| 3 |
+
size 25379244
|
checkpoint-1250/latest
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
global_step1249
|
checkpoint-1250/rng_state_0.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:343f4286599993ed6196c5401702754136843f36dd00127e9d480dcd80159fe3
|
| 3 |
+
size 15984
|
checkpoint-1250/rng_state_1.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f9befffc75eb87b0bc636a17f92325765a402aea6df9004d264ae182c4b3348b
|
| 3 |
+
size 15984
|
checkpoint-1250/rng_state_2.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:71f1a9ba48b7fb3c3e4f1e2cbee76fc776168ccecb572d29900e58fb5dabf633
|
| 3 |
+
size 15984
|
checkpoint-1250/rng_state_3.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e10b5290c005590436e0e89a12ce21b3888b377bb74aaf05b5f52d01666c809d
|
| 3 |
+
size 15984
|
checkpoint-1250/rng_state_4.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:18ad35928e22a674bcbecc7be68aedfab901d49755cd3b1ebaba37e489694d1e
|
| 3 |
+
size 15984
|
checkpoint-1250/rng_state_5.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8fc0855abe8ff42d92eb2c414b25f3a9ba018466841df3ced77b7e05b590df0b
|
| 3 |
+
size 15984
|
checkpoint-1250/rng_state_6.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a40782745bd53ffde678e156f981066d4bc6322666acf70c1894052c32a94fd8
|
| 3 |
+
size 15984
|
checkpoint-1250/rng_state_7.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8622c198dc2441f20ba21ffcfa5ac91baee63fb99abb6137c373932a6870ba75
|
| 3 |
+
size 15984
|
checkpoint-1250/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:85a32e7d79d4bce24e8039d66138f2f6eb81d15d973d067af5d2c4296fc1dd0b
|
| 3 |
+
size 1064
|
checkpoint-1250/special_tokens_map.json
ADDED
|
@@ -0,0 +1,30 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"content": "<s>",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": false,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"eos_token": {
|
| 10 |
+
"content": "<|end|>",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": false,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"pad_token": {
|
| 17 |
+
"content": "<|endoftext|>",
|
| 18 |
+
"lstrip": false,
|
| 19 |
+
"normalized": false,
|
| 20 |
+
"rstrip": false,
|
| 21 |
+
"single_word": false
|
| 22 |
+
},
|
| 23 |
+
"unk_token": {
|
| 24 |
+
"content": "<unk>",
|
| 25 |
+
"lstrip": false,
|
| 26 |
+
"normalized": false,
|
| 27 |
+
"rstrip": false,
|
| 28 |
+
"single_word": false
|
| 29 |
+
}
|
| 30 |
+
}
|
checkpoint-1250/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
checkpoint-1250/tokenizer_config.json
ADDED
|
@@ -0,0 +1,133 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_bos_token": false,
|
| 3 |
+
"add_eos_token": false,
|
| 4 |
+
"add_prefix_space": null,
|
| 5 |
+
"added_tokens_decoder": {
|
| 6 |
+
"0": {
|
| 7 |
+
"content": "<unk>",
|
| 8 |
+
"lstrip": false,
|
| 9 |
+
"normalized": false,
|
| 10 |
+
"rstrip": false,
|
| 11 |
+
"single_word": false,
|
| 12 |
+
"special": true
|
| 13 |
+
},
|
| 14 |
+
"1": {
|
| 15 |
+
"content": "<s>",
|
| 16 |
+
"lstrip": false,
|
| 17 |
+
"normalized": false,
|
| 18 |
+
"rstrip": false,
|
| 19 |
+
"single_word": false,
|
| 20 |
+
"special": true
|
| 21 |
+
},
|
| 22 |
+
"2": {
|
| 23 |
+
"content": "</s>",
|
| 24 |
+
"lstrip": false,
|
| 25 |
+
"normalized": false,
|
| 26 |
+
"rstrip": true,
|
| 27 |
+
"single_word": false,
|
| 28 |
+
"special": false
|
| 29 |
+
},
|
| 30 |
+
"32000": {
|
| 31 |
+
"content": "<|endoftext|>",
|
| 32 |
+
"lstrip": false,
|
| 33 |
+
"normalized": false,
|
| 34 |
+
"rstrip": false,
|
| 35 |
+
"single_word": false,
|
| 36 |
+
"special": true
|
| 37 |
+
},
|
| 38 |
+
"32001": {
|
| 39 |
+
"content": "<|assistant|>",
|
| 40 |
+
"lstrip": false,
|
| 41 |
+
"normalized": false,
|
| 42 |
+
"rstrip": true,
|
| 43 |
+
"single_word": false,
|
| 44 |
+
"special": true
|
| 45 |
+
},
|
| 46 |
+
"32002": {
|
| 47 |
+
"content": "<|placeholder1|>",
|
| 48 |
+
"lstrip": false,
|
| 49 |
+
"normalized": false,
|
| 50 |
+
"rstrip": true,
|
| 51 |
+
"single_word": false,
|
| 52 |
+
"special": true
|
| 53 |
+
},
|
| 54 |
+
"32003": {
|
| 55 |
+
"content": "<|placeholder2|>",
|
| 56 |
+
"lstrip": false,
|
| 57 |
+
"normalized": false,
|
| 58 |
+
"rstrip": true,
|
| 59 |
+
"single_word": false,
|
| 60 |
+
"special": true
|
| 61 |
+
},
|
| 62 |
+
"32004": {
|
| 63 |
+
"content": "<|placeholder3|>",
|
| 64 |
+
"lstrip": false,
|
| 65 |
+
"normalized": false,
|
| 66 |
+
"rstrip": true,
|
| 67 |
+
"single_word": false,
|
| 68 |
+
"special": true
|
| 69 |
+
},
|
| 70 |
+
"32005": {
|
| 71 |
+
"content": "<|placeholder4|>",
|
| 72 |
+
"lstrip": false,
|
| 73 |
+
"normalized": false,
|
| 74 |
+
"rstrip": true,
|
| 75 |
+
"single_word": false,
|
| 76 |
+
"special": true
|
| 77 |
+
},
|
| 78 |
+
"32006": {
|
| 79 |
+
"content": "<|system|>",
|
| 80 |
+
"lstrip": false,
|
| 81 |
+
"normalized": false,
|
| 82 |
+
"rstrip": true,
|
| 83 |
+
"single_word": false,
|
| 84 |
+
"special": true
|
| 85 |
+
},
|
| 86 |
+
"32007": {
|
| 87 |
+
"content": "<|end|>",
|
| 88 |
+
"lstrip": false,
|
| 89 |
+
"normalized": false,
|
| 90 |
+
"rstrip": false,
|
| 91 |
+
"single_word": false,
|
| 92 |
+
"special": true
|
| 93 |
+
},
|
| 94 |
+
"32008": {
|
| 95 |
+
"content": "<|placeholder5|>",
|
| 96 |
+
"lstrip": false,
|
| 97 |
+
"normalized": false,
|
| 98 |
+
"rstrip": true,
|
| 99 |
+
"single_word": false,
|
| 100 |
+
"special": true
|
| 101 |
+
},
|
| 102 |
+
"32009": {
|
| 103 |
+
"content": "<|placeholder6|>",
|
| 104 |
+
"lstrip": false,
|
| 105 |
+
"normalized": false,
|
| 106 |
+
"rstrip": true,
|
| 107 |
+
"single_word": false,
|
| 108 |
+
"special": true
|
| 109 |
+
},
|
| 110 |
+
"32010": {
|
| 111 |
+
"content": "<|user|>",
|
| 112 |
+
"lstrip": false,
|
| 113 |
+
"normalized": false,
|
| 114 |
+
"rstrip": true,
|
| 115 |
+
"single_word": false,
|
| 116 |
+
"special": true
|
| 117 |
+
}
|
| 118 |
+
},
|
| 119 |
+
"bos_token": "<s>",
|
| 120 |
+
"chat_template": "{% set system_message = 'You are a helpful AI assistant.' %}{% if messages[0]['role'] == 'system' %}{% set system_message = messages[0]['content'] %}{% endif %}{% if system_message is defined %}{{ '<s>' + '<|system|>\n' + system_message + '<|end|>\n' }}{% endif %}{% for message in messages %}{% set content = message['content'] %}{% if message['role'] == 'user' %}{{ '<|user|>\n' + content + '<|end|>\n<|assistant|>\n' }}{% elif message['role'] == 'assistant' %}{{ content + '<|end|>' + '\n' }}{% endif %}{% endfor %}",
|
| 121 |
+
"clean_up_tokenization_spaces": false,
|
| 122 |
+
"eos_token": "<|end|>",
|
| 123 |
+
"extra_special_tokens": {},
|
| 124 |
+
"legacy": false,
|
| 125 |
+
"model_max_length": 4096,
|
| 126 |
+
"pad_token": "<|endoftext|>",
|
| 127 |
+
"padding_side": "right",
|
| 128 |
+
"sp_model_kwargs": {},
|
| 129 |
+
"split_special_tokens": false,
|
| 130 |
+
"tokenizer_class": "LlamaTokenizer",
|
| 131 |
+
"unk_token": "<unk>",
|
| 132 |
+
"use_default_system_prompt": false
|
| 133 |
+
}
|
checkpoint-1250/trainer_state.json
ADDED
|
@@ -0,0 +1,2308 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 1.5866666666666667,
|
| 5 |
+
"eval_steps": 50,
|
| 6 |
+
"global_step": 1250,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.012698412698412698,
|
| 13 |
+
"grad_norm": 0.04716634005308151,
|
| 14 |
+
"learning_rate": 4.999451708687114e-06,
|
| 15 |
+
"logits/chosen": 14.883691787719727,
|
| 16 |
+
"logits/rejected": 15.016583442687988,
|
| 17 |
+
"logps/chosen": -0.29512909054756165,
|
| 18 |
+
"logps/rejected": -0.30033987760543823,
|
| 19 |
+
"loss": 1.0007,
|
| 20 |
+
"rewards/accuracies": 0.36250001192092896,
|
| 21 |
+
"rewards/chosen": -0.4426936209201813,
|
| 22 |
+
"rewards/margins": 0.007816222496330738,
|
| 23 |
+
"rewards/rejected": -0.45050984621047974,
|
| 24 |
+
"step": 10
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.025396825396825397,
|
| 28 |
+
"grad_norm": 0.04392225295305252,
|
| 29 |
+
"learning_rate": 4.997807075247147e-06,
|
| 30 |
+
"logits/chosen": 14.705224990844727,
|
| 31 |
+
"logits/rejected": 14.737253189086914,
|
| 32 |
+
"logps/chosen": -0.3201700747013092,
|
| 33 |
+
"logps/rejected": -0.280050128698349,
|
| 34 |
+
"loss": 0.9984,
|
| 35 |
+
"rewards/accuracies": 0.26249998807907104,
|
| 36 |
+
"rewards/chosen": -0.4802550673484802,
|
| 37 |
+
"rewards/margins": -0.060179851949214935,
|
| 38 |
+
"rewards/rejected": -0.4200752377510071,
|
| 39 |
+
"step": 20
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.0380952380952381,
|
| 43 |
+
"grad_norm": 0.04989476501941681,
|
| 44 |
+
"learning_rate": 4.9950668210706795e-06,
|
| 45 |
+
"logits/chosen": 15.147977828979492,
|
| 46 |
+
"logits/rejected": 15.080279350280762,
|
| 47 |
+
"logps/chosen": -0.3157380223274231,
|
| 48 |
+
"logps/rejected": -0.2997627556324005,
|
| 49 |
+
"loss": 0.9991,
|
| 50 |
+
"rewards/accuracies": 0.2874999940395355,
|
| 51 |
+
"rewards/chosen": -0.47360706329345703,
|
| 52 |
+
"rewards/margins": -0.023962898179888725,
|
| 53 |
+
"rewards/rejected": -0.44964417815208435,
|
| 54 |
+
"step": 30
|
| 55 |
+
},
|
| 56 |
+
{
|
| 57 |
+
"epoch": 0.050793650793650794,
|
| 58 |
+
"grad_norm": 0.054729390889406204,
|
| 59 |
+
"learning_rate": 4.9912321481237616e-06,
|
| 60 |
+
"logits/chosen": 15.301165580749512,
|
| 61 |
+
"logits/rejected": 15.276555061340332,
|
| 62 |
+
"logps/chosen": -0.30470195412635803,
|
| 63 |
+
"logps/rejected": -0.29767152667045593,
|
| 64 |
+
"loss": 0.9849,
|
| 65 |
+
"rewards/accuracies": 0.36250001192092896,
|
| 66 |
+
"rewards/chosen": -0.45705294609069824,
|
| 67 |
+
"rewards/margins": -0.010545584373176098,
|
| 68 |
+
"rewards/rejected": -0.44650736451148987,
|
| 69 |
+
"step": 40
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 0.06349206349206349,
|
| 73 |
+
"grad_norm": 0.06035450100898743,
|
| 74 |
+
"learning_rate": 4.986304738420684e-06,
|
| 75 |
+
"logits/chosen": 14.600168228149414,
|
| 76 |
+
"logits/rejected": 14.7944917678833,
|
| 77 |
+
"logps/chosen": -0.32278841733932495,
|
| 78 |
+
"logps/rejected": -0.3014402687549591,
|
| 79 |
+
"loss": 0.9991,
|
| 80 |
+
"rewards/accuracies": 0.3499999940395355,
|
| 81 |
+
"rewards/chosen": -0.48418259620666504,
|
| 82 |
+
"rewards/margins": -0.03202226758003235,
|
| 83 |
+
"rewards/rejected": -0.4521603584289551,
|
| 84 |
+
"step": 50
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 0.06349206349206349,
|
| 88 |
+
"eval_logits/chosen": 15.261337280273438,
|
| 89 |
+
"eval_logits/rejected": 15.51547908782959,
|
| 90 |
+
"eval_logps/chosen": -0.3022651970386505,
|
| 91 |
+
"eval_logps/rejected": -0.3061661124229431,
|
| 92 |
+
"eval_loss": 0.9846106171607971,
|
| 93 |
+
"eval_rewards/accuracies": 0.40625,
|
| 94 |
+
"eval_rewards/chosen": -0.4533977508544922,
|
| 95 |
+
"eval_rewards/margins": 0.005851435009390116,
|
| 96 |
+
"eval_rewards/rejected": -0.45924919843673706,
|
| 97 |
+
"eval_runtime": 18.4033,
|
| 98 |
+
"eval_samples_per_second": 27.712,
|
| 99 |
+
"eval_steps_per_second": 3.478,
|
| 100 |
+
"step": 50
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.0761904761904762,
|
| 104 |
+
"grad_norm": 0.052623528987169266,
|
| 105 |
+
"learning_rate": 4.980286753286196e-06,
|
| 106 |
+
"logits/chosen": 15.2726469039917,
|
| 107 |
+
"logits/rejected": 15.245884895324707,
|
| 108 |
+
"logps/chosen": -0.31394433975219727,
|
| 109 |
+
"logps/rejected": -0.2805258333683014,
|
| 110 |
+
"loss": 1.0063,
|
| 111 |
+
"rewards/accuracies": 0.22499999403953552,
|
| 112 |
+
"rewards/chosen": -0.4709165096282959,
|
| 113 |
+
"rewards/margins": -0.05012776702642441,
|
| 114 |
+
"rewards/rejected": -0.4207887053489685,
|
| 115 |
+
"step": 60
|
| 116 |
+
},
|
| 117 |
+
{
|
| 118 |
+
"epoch": 0.08888888888888889,
|
| 119 |
+
"grad_norm": 0.3512271046638489,
|
| 120 |
+
"learning_rate": 4.973180832407471e-06,
|
| 121 |
+
"logits/chosen": 15.302217483520508,
|
| 122 |
+
"logits/rejected": 15.387911796569824,
|
| 123 |
+
"logps/chosen": -0.30377697944641113,
|
| 124 |
+
"logps/rejected": -0.29276782274246216,
|
| 125 |
+
"loss": 0.9888,
|
| 126 |
+
"rewards/accuracies": 0.3499999940395355,
|
| 127 |
+
"rewards/chosen": -0.4556654393672943,
|
| 128 |
+
"rewards/margins": -0.016513748094439507,
|
| 129 |
+
"rewards/rejected": -0.43915170431137085,
|
| 130 |
+
"step": 70
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 0.10158730158730159,
|
| 134 |
+
"grad_norm": 0.056646134704351425,
|
| 135 |
+
"learning_rate": 4.964990092676263e-06,
|
| 136 |
+
"logits/chosen": 15.843823432922363,
|
| 137 |
+
"logits/rejected": 15.895855903625488,
|
| 138 |
+
"logps/chosen": -0.3091468811035156,
|
| 139 |
+
"logps/rejected": -0.27621737122535706,
|
| 140 |
+
"loss": 0.9926,
|
| 141 |
+
"rewards/accuracies": 0.3375000059604645,
|
| 142 |
+
"rewards/chosen": -0.46372032165527344,
|
| 143 |
+
"rewards/margins": -0.04939427971839905,
|
| 144 |
+
"rewards/rejected": -0.4143260419368744,
|
| 145 |
+
"step": 80
|
| 146 |
+
},
|
| 147 |
+
{
|
| 148 |
+
"epoch": 0.11428571428571428,
|
| 149 |
+
"grad_norm": 0.062348950654268265,
|
| 150 |
+
"learning_rate": 4.9557181268217225e-06,
|
| 151 |
+
"logits/chosen": 15.522564888000488,
|
| 152 |
+
"logits/rejected": 15.576242446899414,
|
| 153 |
+
"logps/chosen": -0.3092329502105713,
|
| 154 |
+
"logps/rejected": -0.27670228481292725,
|
| 155 |
+
"loss": 0.9812,
|
| 156 |
+
"rewards/accuracies": 0.22499999403953552,
|
| 157 |
+
"rewards/chosen": -0.4638494551181793,
|
| 158 |
+
"rewards/margins": -0.048796020448207855,
|
| 159 |
+
"rewards/rejected": -0.41505345702171326,
|
| 160 |
+
"step": 90
|
| 161 |
+
},
|
| 162 |
+
{
|
| 163 |
+
"epoch": 0.12698412698412698,
|
| 164 |
+
"grad_norm": 0.06779928505420685,
|
| 165 |
+
"learning_rate": 4.9453690018345144e-06,
|
| 166 |
+
"logits/chosen": 15.999185562133789,
|
| 167 |
+
"logits/rejected": 15.802484512329102,
|
| 168 |
+
"logps/chosen": -0.3173079490661621,
|
| 169 |
+
"logps/rejected": -0.283602774143219,
|
| 170 |
+
"loss": 0.981,
|
| 171 |
+
"rewards/accuracies": 0.2750000059604645,
|
| 172 |
+
"rewards/chosen": -0.4759618639945984,
|
| 173 |
+
"rewards/margins": -0.050557754933834076,
|
| 174 |
+
"rewards/rejected": -0.4254041314125061,
|
| 175 |
+
"step": 100
|
| 176 |
+
},
|
| 177 |
+
{
|
| 178 |
+
"epoch": 0.12698412698412698,
|
| 179 |
+
"eval_logits/chosen": 15.745854377746582,
|
| 180 |
+
"eval_logits/rejected": 16.0076961517334,
|
| 181 |
+
"eval_logps/chosen": -0.29866182804107666,
|
| 182 |
+
"eval_logps/rejected": -0.3121251165866852,
|
| 183 |
+
"eval_loss": 0.9764130115509033,
|
| 184 |
+
"eval_rewards/accuracies": 0.453125,
|
| 185 |
+
"eval_rewards/chosen": -0.4479926824569702,
|
| 186 |
+
"eval_rewards/margins": 0.02019493840634823,
|
| 187 |
+
"eval_rewards/rejected": -0.4681876599788666,
|
| 188 |
+
"eval_runtime": 18.3383,
|
| 189 |
+
"eval_samples_per_second": 27.811,
|
| 190 |
+
"eval_steps_per_second": 3.49,
|
| 191 |
+
"step": 100
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.13968253968253969,
|
| 195 |
+
"grad_norm": 0.07346926629543304,
|
| 196 |
+
"learning_rate": 4.933947257182901e-06,
|
| 197 |
+
"logits/chosen": 15.723541259765625,
|
| 198 |
+
"logits/rejected": 15.764880180358887,
|
| 199 |
+
"logps/chosen": -0.3212783634662628,
|
| 200 |
+
"logps/rejected": -0.3159862160682678,
|
| 201 |
+
"loss": 0.9859,
|
| 202 |
+
"rewards/accuracies": 0.4124999940395355,
|
| 203 |
+
"rewards/chosen": -0.4819175601005554,
|
| 204 |
+
"rewards/margins": -0.00793826300650835,
|
| 205 |
+
"rewards/rejected": -0.47397929430007935,
|
| 206 |
+
"step": 110
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.1523809523809524,
|
| 210 |
+
"grad_norm": 0.06748441606760025,
|
| 211 |
+
"learning_rate": 4.921457902821578e-06,
|
| 212 |
+
"logits/chosen": 15.817390441894531,
|
| 213 |
+
"logits/rejected": 16.023067474365234,
|
| 214 |
+
"logps/chosen": -0.29811763763427734,
|
| 215 |
+
"logps/rejected": -0.30654552578926086,
|
| 216 |
+
"loss": 0.9827,
|
| 217 |
+
"rewards/accuracies": 0.375,
|
| 218 |
+
"rewards/chosen": -0.447176456451416,
|
| 219 |
+
"rewards/margins": 0.01264181174337864,
|
| 220 |
+
"rewards/rejected": -0.4598182737827301,
|
| 221 |
+
"step": 120
|
| 222 |
+
},
|
| 223 |
+
{
|
| 224 |
+
"epoch": 0.16507936507936508,
|
| 225 |
+
"grad_norm": 0.08042758703231812,
|
| 226 |
+
"learning_rate": 4.907906416994146e-06,
|
| 227 |
+
"logits/chosen": 15.45033073425293,
|
| 228 |
+
"logits/rejected": 15.865328788757324,
|
| 229 |
+
"logps/chosen": -0.27259278297424316,
|
| 230 |
+
"logps/rejected": -0.31338179111480713,
|
| 231 |
+
"loss": 0.9743,
|
| 232 |
+
"rewards/accuracies": 0.48750001192092896,
|
| 233 |
+
"rewards/chosen": -0.40888920426368713,
|
| 234 |
+
"rewards/margins": 0.06118353083729744,
|
| 235 |
+
"rewards/rejected": -0.4700726866722107,
|
| 236 |
+
"step": 130
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 0.17777777777777778,
|
| 240 |
+
"grad_norm": 0.15669430792331696,
|
| 241 |
+
"learning_rate": 4.893298743830168e-06,
|
| 242 |
+
"logits/chosen": 16.17765235900879,
|
| 243 |
+
"logits/rejected": 15.951130867004395,
|
| 244 |
+
"logps/chosen": -0.3015133738517761,
|
| 245 |
+
"logps/rejected": -0.2970424294471741,
|
| 246 |
+
"loss": 0.9824,
|
| 247 |
+
"rewards/accuracies": 0.36250001192092896,
|
| 248 |
+
"rewards/chosen": -0.4522700905799866,
|
| 249 |
+
"rewards/margins": -0.006706444080919027,
|
| 250 |
+
"rewards/rejected": -0.44556355476379395,
|
| 251 |
+
"step": 140
|
| 252 |
+
},
|
| 253 |
+
{
|
| 254 |
+
"epoch": 0.19047619047619047,
|
| 255 |
+
"grad_norm": 0.07115967571735382,
|
| 256 |
+
"learning_rate": 4.8776412907378845e-06,
|
| 257 |
+
"logits/chosen": 15.552728652954102,
|
| 258 |
+
"logits/rejected": 15.788042068481445,
|
| 259 |
+
"logps/chosen": -0.28624147176742554,
|
| 260 |
+
"logps/rejected": -0.3222900331020355,
|
| 261 |
+
"loss": 0.9727,
|
| 262 |
+
"rewards/accuracies": 0.5,
|
| 263 |
+
"rewards/chosen": -0.4293622076511383,
|
| 264 |
+
"rewards/margins": 0.05407290905714035,
|
| 265 |
+
"rewards/rejected": -0.48343515396118164,
|
| 266 |
+
"step": 150
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 0.19047619047619047,
|
| 270 |
+
"eval_logits/chosen": 15.72143268585205,
|
| 271 |
+
"eval_logits/rejected": 15.974818229675293,
|
| 272 |
+
"eval_logps/chosen": -0.29440543055534363,
|
| 273 |
+
"eval_logps/rejected": -0.33279579877853394,
|
| 274 |
+
"eval_loss": 0.9540849328041077,
|
| 275 |
+
"eval_rewards/accuracies": 0.546875,
|
| 276 |
+
"eval_rewards/chosen": -0.44160816073417664,
|
| 277 |
+
"eval_rewards/margins": 0.05758553743362427,
|
| 278 |
+
"eval_rewards/rejected": -0.4991936683654785,
|
| 279 |
+
"eval_runtime": 18.335,
|
| 280 |
+
"eval_samples_per_second": 27.816,
|
| 281 |
+
"eval_steps_per_second": 3.491,
|
| 282 |
+
"step": 150
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 0.20317460317460317,
|
| 286 |
+
"grad_norm": 0.1144075095653534,
|
| 287 |
+
"learning_rate": 4.860940925593703e-06,
|
| 288 |
+
"logits/chosen": 15.63622760772705,
|
| 289 |
+
"logits/rejected": 15.758941650390625,
|
| 290 |
+
"logps/chosen": -0.310846745967865,
|
| 291 |
+
"logps/rejected": -0.35185256600379944,
|
| 292 |
+
"loss": 0.9517,
|
| 293 |
+
"rewards/accuracies": 0.5249999761581421,
|
| 294 |
+
"rewards/chosen": -0.4662701487541199,
|
| 295 |
+
"rewards/margins": 0.061508744955062866,
|
| 296 |
+
"rewards/rejected": -0.5277789235115051,
|
| 297 |
+
"step": 160
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 0.21587301587301588,
|
| 301 |
+
"grad_norm": 0.1173442155122757,
|
| 302 |
+
"learning_rate": 4.84320497372973e-06,
|
| 303 |
+
"logits/chosen": 15.8660306930542,
|
| 304 |
+
"logits/rejected": 15.83642292022705,
|
| 305 |
+
"logps/chosen": -0.279310017824173,
|
| 306 |
+
"logps/rejected": -0.3301998972892761,
|
| 307 |
+
"loss": 0.9439,
|
| 308 |
+
"rewards/accuracies": 0.5,
|
| 309 |
+
"rewards/chosen": -0.41896501183509827,
|
| 310 |
+
"rewards/margins": 0.07633484899997711,
|
| 311 |
+
"rewards/rejected": -0.49529990553855896,
|
| 312 |
+
"step": 170
|
| 313 |
+
},
|
| 314 |
+
{
|
| 315 |
+
"epoch": 0.22857142857142856,
|
| 316 |
+
"grad_norm": 0.204214945435524,
|
| 317 |
+
"learning_rate": 4.824441214720629e-06,
|
| 318 |
+
"logits/chosen": 15.557527542114258,
|
| 319 |
+
"logits/rejected": 15.321504592895508,
|
| 320 |
+
"logps/chosen": -0.3029389977455139,
|
| 321 |
+
"logps/rejected": -0.3163699507713318,
|
| 322 |
+
"loss": 0.9369,
|
| 323 |
+
"rewards/accuracies": 0.38749998807907104,
|
| 324 |
+
"rewards/chosen": -0.4544084668159485,
|
| 325 |
+
"rewards/margins": 0.020146463066339493,
|
| 326 |
+
"rewards/rejected": -0.47455495595932007,
|
| 327 |
+
"step": 180
|
| 328 |
+
},
|
| 329 |
+
{
|
| 330 |
+
"epoch": 0.24126984126984127,
|
| 331 |
+
"grad_norm": 0.14012843370437622,
|
| 332 |
+
"learning_rate": 4.804657878971252e-06,
|
| 333 |
+
"logits/chosen": 15.567469596862793,
|
| 334 |
+
"logits/rejected": 15.717155456542969,
|
| 335 |
+
"logps/chosen": -0.3400501310825348,
|
| 336 |
+
"logps/rejected": -0.3475819230079651,
|
| 337 |
+
"loss": 0.9443,
|
| 338 |
+
"rewards/accuracies": 0.38749998807907104,
|
| 339 |
+
"rewards/chosen": -0.5100752115249634,
|
| 340 |
+
"rewards/margins": 0.011297682300209999,
|
| 341 |
+
"rewards/rejected": -0.5213728547096252,
|
| 342 |
+
"step": 190
|
| 343 |
+
},
|
| 344 |
+
{
|
| 345 |
+
"epoch": 0.25396825396825395,
|
| 346 |
+
"grad_norm": 0.12908300757408142,
|
| 347 |
+
"learning_rate": 4.783863644106502e-06,
|
| 348 |
+
"logits/chosen": 14.985162734985352,
|
| 349 |
+
"logits/rejected": 15.223039627075195,
|
| 350 |
+
"logps/chosen": -0.2876646816730499,
|
| 351 |
+
"logps/rejected": -0.3777172267436981,
|
| 352 |
+
"loss": 0.9034,
|
| 353 |
+
"rewards/accuracies": 0.48750001192092896,
|
| 354 |
+
"rewards/chosen": -0.4314970076084137,
|
| 355 |
+
"rewards/margins": 0.13507887721061707,
|
| 356 |
+
"rewards/rejected": -0.5665758848190308,
|
| 357 |
+
"step": 200
|
| 358 |
+
},
|
| 359 |
+
{
|
| 360 |
+
"epoch": 0.25396825396825395,
|
| 361 |
+
"eval_logits/chosen": 15.242753028869629,
|
| 362 |
+
"eval_logits/rejected": 15.523069381713867,
|
| 363 |
+
"eval_logps/chosen": -0.2961229085922241,
|
| 364 |
+
"eval_logps/rejected": -0.3966684341430664,
|
| 365 |
+
"eval_loss": 0.9075753092765808,
|
| 366 |
+
"eval_rewards/accuracies": 0.59375,
|
| 367 |
+
"eval_rewards/chosen": -0.4441843330860138,
|
| 368 |
+
"eval_rewards/margins": 0.1508183479309082,
|
| 369 |
+
"eval_rewards/rejected": -0.5950026512145996,
|
| 370 |
+
"eval_runtime": 18.341,
|
| 371 |
+
"eval_samples_per_second": 27.807,
|
| 372 |
+
"eval_steps_per_second": 3.489,
|
| 373 |
+
"step": 200
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.26666666666666666,
|
| 377 |
+
"grad_norm": 0.13152211904525757,
|
| 378 |
+
"learning_rate": 4.762067631165049e-06,
|
| 379 |
+
"logits/chosen": 15.283185958862305,
|
| 380 |
+
"logits/rejected": 15.545951843261719,
|
| 381 |
+
"logps/chosen": -0.30045825242996216,
|
| 382 |
+
"logps/rejected": -0.4519944190979004,
|
| 383 |
+
"loss": 0.9085,
|
| 384 |
+
"rewards/accuracies": 0.637499988079071,
|
| 385 |
+
"rewards/chosen": -0.45068734884262085,
|
| 386 |
+
"rewards/margins": 0.22730426490306854,
|
| 387 |
+
"rewards/rejected": -0.6779916286468506,
|
| 388 |
+
"step": 210
|
| 389 |
+
},
|
| 390 |
+
{
|
| 391 |
+
"epoch": 0.27936507936507937,
|
| 392 |
+
"grad_norm": 0.14594660699367523,
|
| 393 |
+
"learning_rate": 4.7392794005985324e-06,
|
| 394 |
+
"logits/chosen": 15.53101921081543,
|
| 395 |
+
"logits/rejected": 15.322232246398926,
|
| 396 |
+
"logps/chosen": -0.3336530327796936,
|
| 397 |
+
"logps/rejected": -0.36337023973464966,
|
| 398 |
+
"loss": 0.9188,
|
| 399 |
+
"rewards/accuracies": 0.44999998807907104,
|
| 400 |
+
"rewards/chosen": -0.5004795789718628,
|
| 401 |
+
"rewards/margins": 0.04457578808069229,
|
| 402 |
+
"rewards/rejected": -0.5450553894042969,
|
| 403 |
+
"step": 220
|
| 404 |
+
},
|
| 405 |
+
{
|
| 406 |
+
"epoch": 0.2920634920634921,
|
| 407 |
+
"grad_norm": 0.14854153990745544,
|
| 408 |
+
"learning_rate": 4.715508948078037e-06,
|
| 409 |
+
"logits/chosen": 14.157003402709961,
|
| 410 |
+
"logits/rejected": 14.34089183807373,
|
| 411 |
+
"logps/chosen": -0.28324443101882935,
|
| 412 |
+
"logps/rejected": -0.41318759322166443,
|
| 413 |
+
"loss": 0.9015,
|
| 414 |
+
"rewards/accuracies": 0.5375000238418579,
|
| 415 |
+
"rewards/chosen": -0.4248666763305664,
|
| 416 |
+
"rewards/margins": 0.19491472840309143,
|
| 417 |
+
"rewards/rejected": -0.6197813749313354,
|
| 418 |
+
"step": 230
|
| 419 |
+
},
|
| 420 |
+
{
|
| 421 |
+
"epoch": 0.3047619047619048,
|
| 422 |
+
"grad_norm": 0.19532087445259094,
|
| 423 |
+
"learning_rate": 4.690766700109659e-06,
|
| 424 |
+
"logits/chosen": 14.729398727416992,
|
| 425 |
+
"logits/rejected": 14.97203254699707,
|
| 426 |
+
"logps/chosen": -0.3013172149658203,
|
| 427 |
+
"logps/rejected": -0.41750937700271606,
|
| 428 |
+
"loss": 0.8814,
|
| 429 |
+
"rewards/accuracies": 0.512499988079071,
|
| 430 |
+
"rewards/chosen": -0.45197582244873047,
|
| 431 |
+
"rewards/margins": 0.17428824305534363,
|
| 432 |
+
"rewards/rejected": -0.6262640357017517,
|
| 433 |
+
"step": 240
|
| 434 |
+
},
|
| 435 |
+
{
|
| 436 |
+
"epoch": 0.31746031746031744,
|
| 437 |
+
"grad_norm": 0.5795227885246277,
|
| 438 |
+
"learning_rate": 4.665063509461098e-06,
|
| 439 |
+
"logits/chosen": 14.328412055969238,
|
| 440 |
+
"logits/rejected": 14.645428657531738,
|
| 441 |
+
"logps/chosen": -0.3055194914340973,
|
| 442 |
+
"logps/rejected": -0.509621262550354,
|
| 443 |
+
"loss": 0.8478,
|
| 444 |
+
"rewards/accuracies": 0.512499988079071,
|
| 445 |
+
"rewards/chosen": -0.45827922224998474,
|
| 446 |
+
"rewards/margins": 0.30615273118019104,
|
| 447 |
+
"rewards/rejected": -0.764431893825531,
|
| 448 |
+
"step": 250
|
| 449 |
+
},
|
| 450 |
+
{
|
| 451 |
+
"epoch": 0.31746031746031744,
|
| 452 |
+
"eval_logits/chosen": 14.310781478881836,
|
| 453 |
+
"eval_logits/rejected": 14.5652494430542,
|
| 454 |
+
"eval_logps/chosen": -0.3155084252357483,
|
| 455 |
+
"eval_logps/rejected": -0.8301784992218018,
|
| 456 |
+
"eval_loss": 0.7700436115264893,
|
| 457 |
+
"eval_rewards/accuracies": 0.640625,
|
| 458 |
+
"eval_rewards/chosen": -0.47326260805130005,
|
| 459 |
+
"eval_rewards/margins": 0.7720052003860474,
|
| 460 |
+
"eval_rewards/rejected": -1.2452678680419922,
|
| 461 |
+
"eval_runtime": 18.3413,
|
| 462 |
+
"eval_samples_per_second": 27.806,
|
| 463 |
+
"eval_steps_per_second": 3.489,
|
| 464 |
+
"step": 250
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.33015873015873015,
|
| 468 |
+
"grad_norm": 0.24876689910888672,
|
| 469 |
+
"learning_rate": 4.638410650401267e-06,
|
| 470 |
+
"logits/chosen": 14.115270614624023,
|
| 471 |
+
"logits/rejected": 14.167577743530273,
|
| 472 |
+
"logps/chosen": -0.32328444719314575,
|
| 473 |
+
"logps/rejected": -0.8564422726631165,
|
| 474 |
+
"loss": 0.795,
|
| 475 |
+
"rewards/accuracies": 0.5375000238418579,
|
| 476 |
+
"rewards/chosen": -0.484926700592041,
|
| 477 |
+
"rewards/margins": 0.7997367978096008,
|
| 478 |
+
"rewards/rejected": -1.284663438796997,
|
| 479 |
+
"step": 260
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 0.34285714285714286,
|
| 483 |
+
"grad_norm": 0.5779634118080139,
|
| 484 |
+
"learning_rate": 4.610819813755038e-06,
|
| 485 |
+
"logits/chosen": 14.019407272338867,
|
| 486 |
+
"logits/rejected": 13.600751876831055,
|
| 487 |
+
"logps/chosen": -0.35225847363471985,
|
| 488 |
+
"logps/rejected": -1.1924464702606201,
|
| 489 |
+
"loss": 0.7644,
|
| 490 |
+
"rewards/accuracies": 0.5249999761581421,
|
| 491 |
+
"rewards/chosen": -0.5283876657485962,
|
| 492 |
+
"rewards/margins": 1.2602821588516235,
|
| 493 |
+
"rewards/rejected": -1.7886695861816406,
|
| 494 |
+
"step": 270
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 0.35555555555555557,
|
| 498 |
+
"grad_norm": 0.2085695117712021,
|
| 499 |
+
"learning_rate": 4.582303101775249e-06,
|
| 500 |
+
"logits/chosen": 13.448820114135742,
|
| 501 |
+
"logits/rejected": 13.464624404907227,
|
| 502 |
+
"logps/chosen": -0.3442023992538452,
|
| 503 |
+
"logps/rejected": -1.4783053398132324,
|
| 504 |
+
"loss": 0.7671,
|
| 505 |
+
"rewards/accuracies": 0.48750001192092896,
|
| 506 |
+
"rewards/chosen": -0.516303539276123,
|
| 507 |
+
"rewards/margins": 1.7011544704437256,
|
| 508 |
+
"rewards/rejected": -2.2174580097198486,
|
| 509 |
+
"step": 280
|
| 510 |
+
},
|
| 511 |
+
{
|
| 512 |
+
"epoch": 0.3682539682539683,
|
| 513 |
+
"grad_norm": 0.2825860381126404,
|
| 514 |
+
"learning_rate": 4.55287302283426e-06,
|
| 515 |
+
"logits/chosen": 13.038851737976074,
|
| 516 |
+
"logits/rejected": 13.013893127441406,
|
| 517 |
+
"logps/chosen": -0.3524690568447113,
|
| 518 |
+
"logps/rejected": -1.2354861497879028,
|
| 519 |
+
"loss": 0.7922,
|
| 520 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 521 |
+
"rewards/chosen": -0.5287035703659058,
|
| 522 |
+
"rewards/margins": 1.3245255947113037,
|
| 523 |
+
"rewards/rejected": -1.8532291650772095,
|
| 524 |
+
"step": 290
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 0.38095238095238093,
|
| 528 |
+
"grad_norm": 0.30771639943122864,
|
| 529 |
+
"learning_rate": 4.522542485937369e-06,
|
| 530 |
+
"logits/chosen": 13.67170524597168,
|
| 531 |
+
"logits/rejected": 13.185449600219727,
|
| 532 |
+
"logps/chosen": -0.3915162682533264,
|
| 533 |
+
"logps/rejected": -1.3473880290985107,
|
| 534 |
+
"loss": 0.7506,
|
| 535 |
+
"rewards/accuracies": 0.48750001192092896,
|
| 536 |
+
"rewards/chosen": -0.587274432182312,
|
| 537 |
+
"rewards/margins": 1.433807611465454,
|
| 538 |
+
"rewards/rejected": -2.0210821628570557,
|
| 539 |
+
"step": 300
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"epoch": 0.38095238095238093,
|
| 543 |
+
"eval_logits/chosen": 13.198552131652832,
|
| 544 |
+
"eval_logits/rejected": 13.458879470825195,
|
| 545 |
+
"eval_logps/chosen": -0.37132135033607483,
|
| 546 |
+
"eval_logps/rejected": -1.4595866203308105,
|
| 547 |
+
"eval_loss": 0.7391816973686218,
|
| 548 |
+
"eval_rewards/accuracies": 0.640625,
|
| 549 |
+
"eval_rewards/chosen": -0.5569820404052734,
|
| 550 |
+
"eval_rewards/margins": 1.6323981285095215,
|
| 551 |
+
"eval_rewards/rejected": -2.189380168914795,
|
| 552 |
+
"eval_runtime": 18.3117,
|
| 553 |
+
"eval_samples_per_second": 27.851,
|
| 554 |
+
"eval_steps_per_second": 3.495,
|
| 555 |
+
"step": 300
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.39365079365079364,
|
| 559 |
+
"grad_norm": 0.3021286129951477,
|
| 560 |
+
"learning_rate": 4.491324795060491e-06,
|
| 561 |
+
"logits/chosen": 13.347516059875488,
|
| 562 |
+
"logits/rejected": 13.285726547241211,
|
| 563 |
+
"logps/chosen": -0.4108191132545471,
|
| 564 |
+
"logps/rejected": -1.436702013015747,
|
| 565 |
+
"loss": 0.7134,
|
| 566 |
+
"rewards/accuracies": 0.550000011920929,
|
| 567 |
+
"rewards/chosen": -0.6162286400794983,
|
| 568 |
+
"rewards/margins": 1.5388243198394775,
|
| 569 |
+
"rewards/rejected": -2.15505313873291,
|
| 570 |
+
"step": 310
|
| 571 |
+
},
|
| 572 |
+
{
|
| 573 |
+
"epoch": 0.40634920634920635,
|
| 574 |
+
"grad_norm": 0.40123897790908813,
|
| 575 |
+
"learning_rate": 4.4592336433146e-06,
|
| 576 |
+
"logits/chosen": 12.949071884155273,
|
| 577 |
+
"logits/rejected": 12.86626148223877,
|
| 578 |
+
"logps/chosen": -0.42898029088974,
|
| 579 |
+
"logps/rejected": -1.1741807460784912,
|
| 580 |
+
"loss": 0.74,
|
| 581 |
+
"rewards/accuracies": 0.5375000238418579,
|
| 582 |
+
"rewards/chosen": -0.6434704065322876,
|
| 583 |
+
"rewards/margins": 1.1178009510040283,
|
| 584 |
+
"rewards/rejected": -1.7612712383270264,
|
| 585 |
+
"step": 320
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 0.41904761904761906,
|
| 589 |
+
"grad_norm": 0.4613189697265625,
|
| 590 |
+
"learning_rate": 4.426283106939474e-06,
|
| 591 |
+
"logits/chosen": 12.796838760375977,
|
| 592 |
+
"logits/rejected": 12.524165153503418,
|
| 593 |
+
"logps/chosen": -0.48975634574890137,
|
| 594 |
+
"logps/rejected": -1.2766364812850952,
|
| 595 |
+
"loss": 0.7408,
|
| 596 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 597 |
+
"rewards/chosen": -0.7346345782279968,
|
| 598 |
+
"rewards/margins": 1.180320143699646,
|
| 599 |
+
"rewards/rejected": -1.9149547815322876,
|
| 600 |
+
"step": 330
|
| 601 |
+
},
|
| 602 |
+
{
|
| 603 |
+
"epoch": 0.43174603174603177,
|
| 604 |
+
"grad_norm": 0.49141305685043335,
|
| 605 |
+
"learning_rate": 4.3924876391293915e-06,
|
| 606 |
+
"logits/chosen": 12.58080005645752,
|
| 607 |
+
"logits/rejected": 12.5891695022583,
|
| 608 |
+
"logps/chosen": -0.5303409099578857,
|
| 609 |
+
"logps/rejected": -1.6262977123260498,
|
| 610 |
+
"loss": 0.7501,
|
| 611 |
+
"rewards/accuracies": 0.7124999761581421,
|
| 612 |
+
"rewards/chosen": -0.7955113649368286,
|
| 613 |
+
"rewards/margins": 1.643935203552246,
|
| 614 |
+
"rewards/rejected": -2.4394466876983643,
|
| 615 |
+
"step": 340
|
| 616 |
+
},
|
| 617 |
+
{
|
| 618 |
+
"epoch": 0.4444444444444444,
|
| 619 |
+
"grad_norm": 0.7533912062644958,
|
| 620 |
+
"learning_rate": 4.357862063693486e-06,
|
| 621 |
+
"logits/chosen": 12.15321159362793,
|
| 622 |
+
"logits/rejected": 12.38386344909668,
|
| 623 |
+
"logps/chosen": -0.5782582759857178,
|
| 624 |
+
"logps/rejected": -1.580635905265808,
|
| 625 |
+
"loss": 0.7405,
|
| 626 |
+
"rewards/accuracies": 0.699999988079071,
|
| 627 |
+
"rewards/chosen": -0.8673874139785767,
|
| 628 |
+
"rewards/margins": 1.5035666227340698,
|
| 629 |
+
"rewards/rejected": -2.3709540367126465,
|
| 630 |
+
"step": 350
|
| 631 |
+
},
|
| 632 |
+
{
|
| 633 |
+
"epoch": 0.4444444444444444,
|
| 634 |
+
"eval_logits/chosen": 12.111034393310547,
|
| 635 |
+
"eval_logits/rejected": 12.33246898651123,
|
| 636 |
+
"eval_logps/chosen": -0.5941734910011292,
|
| 637 |
+
"eval_logps/rejected": -1.6774102449417114,
|
| 638 |
+
"eval_loss": 0.6953701972961426,
|
| 639 |
+
"eval_rewards/accuracies": 0.765625,
|
| 640 |
+
"eval_rewards/chosen": -0.8912601470947266,
|
| 641 |
+
"eval_rewards/margins": 1.6248550415039062,
|
| 642 |
+
"eval_rewards/rejected": -2.516115188598633,
|
| 643 |
+
"eval_runtime": 18.3366,
|
| 644 |
+
"eval_samples_per_second": 27.813,
|
| 645 |
+
"eval_steps_per_second": 3.49,
|
| 646 |
+
"step": 350
|
| 647 |
+
},
|
| 648 |
+
{
|
| 649 |
+
"epoch": 0.45714285714285713,
|
| 650 |
+
"grad_norm": 1.168564796447754,
|
| 651 |
+
"learning_rate": 4.322421568553529e-06,
|
| 652 |
+
"logits/chosen": 11.706710815429688,
|
| 653 |
+
"logits/rejected": 11.451011657714844,
|
| 654 |
+
"logps/chosen": -0.758882462978363,
|
| 655 |
+
"logps/rejected": -1.842167615890503,
|
| 656 |
+
"loss": 0.702,
|
| 657 |
+
"rewards/accuracies": 0.7124999761581421,
|
| 658 |
+
"rewards/chosen": -1.1383236646652222,
|
| 659 |
+
"rewards/margins": 1.6249277591705322,
|
| 660 |
+
"rewards/rejected": -2.763251304626465,
|
| 661 |
+
"step": 360
|
| 662 |
+
},
|
| 663 |
+
{
|
| 664 |
+
"epoch": 0.46984126984126984,
|
| 665 |
+
"grad_norm": 1.8962547779083252,
|
| 666 |
+
"learning_rate": 4.286181699082008e-06,
|
| 667 |
+
"logits/chosen": 11.41061019897461,
|
| 668 |
+
"logits/rejected": 11.534398078918457,
|
| 669 |
+
"logps/chosen": -1.1785552501678467,
|
| 670 |
+
"logps/rejected": -2.235652446746826,
|
| 671 |
+
"loss": 0.6781,
|
| 672 |
+
"rewards/accuracies": 0.800000011920929,
|
| 673 |
+
"rewards/chosen": -1.7678327560424805,
|
| 674 |
+
"rewards/margins": 1.5856454372406006,
|
| 675 |
+
"rewards/rejected": -3.3534786701202393,
|
| 676 |
+
"step": 370
|
| 677 |
+
},
|
| 678 |
+
{
|
| 679 |
+
"epoch": 0.48253968253968255,
|
| 680 |
+
"grad_norm": 2.9152028560638428,
|
| 681 |
+
"learning_rate": 4.249158351283414e-06,
|
| 682 |
+
"logits/chosen": 11.582561492919922,
|
| 683 |
+
"logits/rejected": 11.439531326293945,
|
| 684 |
+
"logps/chosen": -1.9169971942901611,
|
| 685 |
+
"logps/rejected": -2.788179397583008,
|
| 686 |
+
"loss": 0.6033,
|
| 687 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 688 |
+
"rewards/chosen": -2.875495672225952,
|
| 689 |
+
"rewards/margins": 1.306773066520691,
|
| 690 |
+
"rewards/rejected": -4.182269096374512,
|
| 691 |
+
"step": 380
|
| 692 |
+
},
|
| 693 |
+
{
|
| 694 |
+
"epoch": 0.49523809523809526,
|
| 695 |
+
"grad_norm": 1.5070641040802002,
|
| 696 |
+
"learning_rate": 4.211367764821722e-06,
|
| 697 |
+
"logits/chosen": 11.340807914733887,
|
| 698 |
+
"logits/rejected": 10.927321434020996,
|
| 699 |
+
"logps/chosen": -2.1075730323791504,
|
| 700 |
+
"logps/rejected": -3.4216556549072266,
|
| 701 |
+
"loss": 0.5698,
|
| 702 |
+
"rewards/accuracies": 0.949999988079071,
|
| 703 |
+
"rewards/chosen": -3.1613595485687256,
|
| 704 |
+
"rewards/margins": 1.9711239337921143,
|
| 705 |
+
"rewards/rejected": -5.13248348236084,
|
| 706 |
+
"step": 390
|
| 707 |
+
},
|
| 708 |
+
{
|
| 709 |
+
"epoch": 0.5079365079365079,
|
| 710 |
+
"grad_norm": 2.6913516521453857,
|
| 711 |
+
"learning_rate": 4.172826515897146e-06,
|
| 712 |
+
"logits/chosen": 10.273612976074219,
|
| 713 |
+
"logits/rejected": 10.116204261779785,
|
| 714 |
+
"logps/chosen": -3.0047056674957275,
|
| 715 |
+
"logps/rejected": -4.053646564483643,
|
| 716 |
+
"loss": 0.5869,
|
| 717 |
+
"rewards/accuracies": 0.875,
|
| 718 |
+
"rewards/chosen": -4.507058143615723,
|
| 719 |
+
"rewards/margins": 1.5734113454818726,
|
| 720 |
+
"rewards/rejected": -6.080469608306885,
|
| 721 |
+
"step": 400
|
| 722 |
+
},
|
| 723 |
+
{
|
| 724 |
+
"epoch": 0.5079365079365079,
|
| 725 |
+
"eval_logits/chosen": 9.58273696899414,
|
| 726 |
+
"eval_logits/rejected": 9.765192031860352,
|
| 727 |
+
"eval_logps/chosen": -2.4895575046539307,
|
| 728 |
+
"eval_logps/rejected": -3.8314545154571533,
|
| 729 |
+
"eval_loss": 0.5232856273651123,
|
| 730 |
+
"eval_rewards/accuracies": 0.875,
|
| 731 |
+
"eval_rewards/chosen": -3.7343361377716064,
|
| 732 |
+
"eval_rewards/margins": 2.012845754623413,
|
| 733 |
+
"eval_rewards/rejected": -5.7471818923950195,
|
| 734 |
+
"eval_runtime": 18.3437,
|
| 735 |
+
"eval_samples_per_second": 27.802,
|
| 736 |
+
"eval_steps_per_second": 3.489,
|
| 737 |
+
"step": 400
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 0.5206349206349207,
|
| 741 |
+
"grad_norm": 3.7197885513305664,
|
| 742 |
+
"learning_rate": 4.133551509975264e-06,
|
| 743 |
+
"logits/chosen": 9.081324577331543,
|
| 744 |
+
"logits/rejected": 9.309822082519531,
|
| 745 |
+
"logps/chosen": -2.2554564476013184,
|
| 746 |
+
"logps/rejected": -3.2624363899230957,
|
| 747 |
+
"loss": 0.5774,
|
| 748 |
+
"rewards/accuracies": 0.887499988079071,
|
| 749 |
+
"rewards/chosen": -3.3831849098205566,
|
| 750 |
+
"rewards/margins": 1.5104694366455078,
|
| 751 |
+
"rewards/rejected": -4.893653869628906,
|
| 752 |
+
"step": 410
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 0.5333333333333333,
|
| 756 |
+
"grad_norm": 2.757873773574829,
|
| 757 |
+
"learning_rate": 4.093559974371725e-06,
|
| 758 |
+
"logits/chosen": 8.145927429199219,
|
| 759 |
+
"logits/rejected": 8.402036666870117,
|
| 760 |
+
"logps/chosen": -3.1110777854919434,
|
| 761 |
+
"logps/rejected": -4.202320098876953,
|
| 762 |
+
"loss": 0.5084,
|
| 763 |
+
"rewards/accuracies": 0.862500011920929,
|
| 764 |
+
"rewards/chosen": -4.666616916656494,
|
| 765 |
+
"rewards/margins": 1.636863350868225,
|
| 766 |
+
"rewards/rejected": -6.30348014831543,
|
| 767 |
+
"step": 420
|
| 768 |
+
},
|
| 769 |
+
{
|
| 770 |
+
"epoch": 0.546031746031746,
|
| 771 |
+
"grad_norm": 3.3764102458953857,
|
| 772 |
+
"learning_rate": 4.052869450695776e-06,
|
| 773 |
+
"logits/chosen": 8.132574081420898,
|
| 774 |
+
"logits/rejected": 8.386279106140137,
|
| 775 |
+
"logps/chosen": -3.0871124267578125,
|
| 776 |
+
"logps/rejected": -4.399088382720947,
|
| 777 |
+
"loss": 0.5233,
|
| 778 |
+
"rewards/accuracies": 0.875,
|
| 779 |
+
"rewards/chosen": -4.630668640136719,
|
| 780 |
+
"rewards/margins": 1.967963457107544,
|
| 781 |
+
"rewards/rejected": -6.598631858825684,
|
| 782 |
+
"step": 430
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 0.5587301587301587,
|
| 786 |
+
"grad_norm": 1.8188632726669312,
|
| 787 |
+
"learning_rate": 4.011497787155938e-06,
|
| 788 |
+
"logits/chosen": 8.104793548583984,
|
| 789 |
+
"logits/rejected": 8.25890064239502,
|
| 790 |
+
"logps/chosen": -3.598665237426758,
|
| 791 |
+
"logps/rejected": -5.2809929847717285,
|
| 792 |
+
"loss": 0.5272,
|
| 793 |
+
"rewards/accuracies": 0.824999988079071,
|
| 794 |
+
"rewards/chosen": -5.397997856140137,
|
| 795 |
+
"rewards/margins": 2.5234923362731934,
|
| 796 |
+
"rewards/rejected": -7.921489715576172,
|
| 797 |
+
"step": 440
|
| 798 |
+
},
|
| 799 |
+
{
|
| 800 |
+
"epoch": 0.5714285714285714,
|
| 801 |
+
"grad_norm": 2.528883934020996,
|
| 802 |
+
"learning_rate": 3.969463130731183e-06,
|
| 803 |
+
"logits/chosen": 8.675338745117188,
|
| 804 |
+
"logits/rejected": 8.550267219543457,
|
| 805 |
+
"logps/chosen": -3.582223892211914,
|
| 806 |
+
"logps/rejected": -5.077345848083496,
|
| 807 |
+
"loss": 0.5717,
|
| 808 |
+
"rewards/accuracies": 0.800000011920929,
|
| 809 |
+
"rewards/chosen": -5.373335838317871,
|
| 810 |
+
"rewards/margins": 2.2426836490631104,
|
| 811 |
+
"rewards/rejected": -7.616019248962402,
|
| 812 |
+
"step": 450
|
| 813 |
+
},
|
| 814 |
+
{
|
| 815 |
+
"epoch": 0.5714285714285714,
|
| 816 |
+
"eval_logits/chosen": 8.125189781188965,
|
| 817 |
+
"eval_logits/rejected": 8.239119529724121,
|
| 818 |
+
"eval_logps/chosen": -3.107839584350586,
|
| 819 |
+
"eval_logps/rejected": -4.726868629455566,
|
| 820 |
+
"eval_loss": 0.4568406939506531,
|
| 821 |
+
"eval_rewards/accuracies": 0.875,
|
| 822 |
+
"eval_rewards/chosen": -4.661759376525879,
|
| 823 |
+
"eval_rewards/margins": 2.4285435676574707,
|
| 824 |
+
"eval_rewards/rejected": -7.090303421020508,
|
| 825 |
+
"eval_runtime": 18.3222,
|
| 826 |
+
"eval_samples_per_second": 27.835,
|
| 827 |
+
"eval_steps_per_second": 3.493,
|
| 828 |
+
"step": 450
|
| 829 |
+
},
|
| 830 |
+
{
|
| 831 |
+
"epoch": 0.5841269841269842,
|
| 832 |
+
"grad_norm": 2.5079057216644287,
|
| 833 |
+
"learning_rate": 3.92678391921108e-06,
|
| 834 |
+
"logits/chosen": 8.503251075744629,
|
| 835 |
+
"logits/rejected": 8.457250595092773,
|
| 836 |
+
"logps/chosen": -3.499116897583008,
|
| 837 |
+
"logps/rejected": -5.12787389755249,
|
| 838 |
+
"loss": 0.4803,
|
| 839 |
+
"rewards/accuracies": 0.887499988079071,
|
| 840 |
+
"rewards/chosen": -5.2486748695373535,
|
| 841 |
+
"rewards/margins": 2.4431352615356445,
|
| 842 |
+
"rewards/rejected": -7.69180965423584,
|
| 843 |
+
"step": 460
|
| 844 |
+
},
|
| 845 |
+
{
|
| 846 |
+
"epoch": 0.5968253968253968,
|
| 847 |
+
"grad_norm": 1.869667887687683,
|
| 848 |
+
"learning_rate": 3.88347887310836e-06,
|
| 849 |
+
"logits/chosen": 7.972044467926025,
|
| 850 |
+
"logits/rejected": 7.825163841247559,
|
| 851 |
+
"logps/chosen": -3.6900856494903564,
|
| 852 |
+
"logps/rejected": -5.166090488433838,
|
| 853 |
+
"loss": 0.4733,
|
| 854 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 855 |
+
"rewards/chosen": -5.535128593444824,
|
| 856 |
+
"rewards/margins": 2.2140071392059326,
|
| 857 |
+
"rewards/rejected": -7.749135494232178,
|
| 858 |
+
"step": 470
|
| 859 |
+
},
|
| 860 |
+
{
|
| 861 |
+
"epoch": 0.6095238095238096,
|
| 862 |
+
"grad_norm": 2.8777735233306885,
|
| 863 |
+
"learning_rate": 3.839566987447492e-06,
|
| 864 |
+
"logits/chosen": 8.073511123657227,
|
| 865 |
+
"logits/rejected": 8.345193862915039,
|
| 866 |
+
"logps/chosen": -3.7006676197052,
|
| 867 |
+
"logps/rejected": -5.271113395690918,
|
| 868 |
+
"loss": 0.5048,
|
| 869 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 870 |
+
"rewards/chosen": -5.551001071929932,
|
| 871 |
+
"rewards/margins": 2.3556694984436035,
|
| 872 |
+
"rewards/rejected": -7.906670570373535,
|
| 873 |
+
"step": 480
|
| 874 |
+
},
|
| 875 |
+
{
|
| 876 |
+
"epoch": 0.6222222222222222,
|
| 877 |
+
"grad_norm": 2.7042839527130127,
|
| 878 |
+
"learning_rate": 3.795067523432826e-06,
|
| 879 |
+
"logits/chosen": 8.602261543273926,
|
| 880 |
+
"logits/rejected": 8.32430362701416,
|
| 881 |
+
"logps/chosen": -3.9612879753112793,
|
| 882 |
+
"logps/rejected": -6.123431205749512,
|
| 883 |
+
"loss": 0.4728,
|
| 884 |
+
"rewards/accuracies": 0.949999988079071,
|
| 885 |
+
"rewards/chosen": -5.94193172454834,
|
| 886 |
+
"rewards/margins": 3.2432150840759277,
|
| 887 |
+
"rewards/rejected": -9.185147285461426,
|
| 888 |
+
"step": 490
|
| 889 |
+
},
|
| 890 |
+
{
|
| 891 |
+
"epoch": 0.6349206349206349,
|
| 892 |
+
"grad_norm": 2.1625821590423584,
|
| 893 |
+
"learning_rate": 3.7500000000000005e-06,
|
| 894 |
+
"logits/chosen": 7.857954978942871,
|
| 895 |
+
"logits/rejected": 7.460604190826416,
|
| 896 |
+
"logps/chosen": -3.581023693084717,
|
| 897 |
+
"logps/rejected": -5.477323532104492,
|
| 898 |
+
"loss": 0.4296,
|
| 899 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 900 |
+
"rewards/chosen": -5.371535778045654,
|
| 901 |
+
"rewards/margins": 2.844449520111084,
|
| 902 |
+
"rewards/rejected": -8.215986251831055,
|
| 903 |
+
"step": 500
|
| 904 |
+
},
|
| 905 |
+
{
|
| 906 |
+
"epoch": 0.6349206349206349,
|
| 907 |
+
"eval_logits/chosen": 7.520671844482422,
|
| 908 |
+
"eval_logits/rejected": 7.586723327636719,
|
| 909 |
+
"eval_logps/chosen": -3.1659929752349854,
|
| 910 |
+
"eval_logps/rejected": -4.955198764801025,
|
| 911 |
+
"eval_loss": 0.4266711175441742,
|
| 912 |
+
"eval_rewards/accuracies": 0.890625,
|
| 913 |
+
"eval_rewards/chosen": -4.748989105224609,
|
| 914 |
+
"eval_rewards/margins": 2.683809518814087,
|
| 915 |
+
"eval_rewards/rejected": -7.432799339294434,
|
| 916 |
+
"eval_runtime": 18.3456,
|
| 917 |
+
"eval_samples_per_second": 27.8,
|
| 918 |
+
"eval_steps_per_second": 3.489,
|
| 919 |
+
"step": 500
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"epoch": 0.6476190476190476,
|
| 923 |
+
"grad_norm": 4.2689104080200195,
|
| 924 |
+
"learning_rate": 3.7043841852542884e-06,
|
| 925 |
+
"logits/chosen": 7.948687553405762,
|
| 926 |
+
"logits/rejected": 7.656899929046631,
|
| 927 |
+
"logps/chosen": -2.8974719047546387,
|
| 928 |
+
"logps/rejected": -4.501969337463379,
|
| 929 |
+
"loss": 0.4981,
|
| 930 |
+
"rewards/accuracies": 0.875,
|
| 931 |
+
"rewards/chosen": -4.346207618713379,
|
| 932 |
+
"rewards/margins": 2.4067459106445312,
|
| 933 |
+
"rewards/rejected": -6.752954006195068,
|
| 934 |
+
"step": 510
|
| 935 |
+
},
|
| 936 |
+
{
|
| 937 |
+
"epoch": 0.6603174603174603,
|
| 938 |
+
"grad_norm": 3.299290895462036,
|
| 939 |
+
"learning_rate": 3.658240087799655e-06,
|
| 940 |
+
"logits/chosen": 7.676476955413818,
|
| 941 |
+
"logits/rejected": 7.332755088806152,
|
| 942 |
+
"logps/chosen": -3.9289379119873047,
|
| 943 |
+
"logps/rejected": -5.498073577880859,
|
| 944 |
+
"loss": 0.4736,
|
| 945 |
+
"rewards/accuracies": 0.8374999761581421,
|
| 946 |
+
"rewards/chosen": -5.893406867980957,
|
| 947 |
+
"rewards/margins": 2.353703498840332,
|
| 948 |
+
"rewards/rejected": -8.247110366821289,
|
| 949 |
+
"step": 520
|
| 950 |
+
},
|
| 951 |
+
{
|
| 952 |
+
"epoch": 0.6730158730158731,
|
| 953 |
+
"grad_norm": 3.1058828830718994,
|
| 954 |
+
"learning_rate": 3.611587947962319e-06,
|
| 955 |
+
"logits/chosen": 7.516615867614746,
|
| 956 |
+
"logits/rejected": 7.344383239746094,
|
| 957 |
+
"logps/chosen": -3.57861328125,
|
| 958 |
+
"logps/rejected": -5.48795223236084,
|
| 959 |
+
"loss": 0.502,
|
| 960 |
+
"rewards/accuracies": 0.887499988079071,
|
| 961 |
+
"rewards/chosen": -5.367920398712158,
|
| 962 |
+
"rewards/margins": 2.8640081882476807,
|
| 963 |
+
"rewards/rejected": -8.231928825378418,
|
| 964 |
+
"step": 530
|
| 965 |
+
},
|
| 966 |
+
{
|
| 967 |
+
"epoch": 0.6857142857142857,
|
| 968 |
+
"grad_norm": 3.874868869781494,
|
| 969 |
+
"learning_rate": 3.564448228912682e-06,
|
| 970 |
+
"logits/chosen": 7.619947910308838,
|
| 971 |
+
"logits/rejected": 7.048402309417725,
|
| 972 |
+
"logps/chosen": -4.164804935455322,
|
| 973 |
+
"logps/rejected": -5.771735191345215,
|
| 974 |
+
"loss": 0.44,
|
| 975 |
+
"rewards/accuracies": 0.925000011920929,
|
| 976 |
+
"rewards/chosen": -6.247206687927246,
|
| 977 |
+
"rewards/margins": 2.4103951454162598,
|
| 978 |
+
"rewards/rejected": -8.657602310180664,
|
| 979 |
+
"step": 540
|
| 980 |
+
},
|
| 981 |
+
{
|
| 982 |
+
"epoch": 0.6984126984126984,
|
| 983 |
+
"grad_norm": 1.7583000659942627,
|
| 984 |
+
"learning_rate": 3.516841607689501e-06,
|
| 985 |
+
"logits/chosen": 7.068112850189209,
|
| 986 |
+
"logits/rejected": 7.042977333068848,
|
| 987 |
+
"logps/chosen": -3.760105609893799,
|
| 988 |
+
"logps/rejected": -5.3382248878479,
|
| 989 |
+
"loss": 0.4751,
|
| 990 |
+
"rewards/accuracies": 0.887499988079071,
|
| 991 |
+
"rewards/chosen": -5.640158653259277,
|
| 992 |
+
"rewards/margins": 2.3671793937683105,
|
| 993 |
+
"rewards/rejected": -8.00733757019043,
|
| 994 |
+
"step": 550
|
| 995 |
+
},
|
| 996 |
+
{
|
| 997 |
+
"epoch": 0.6984126984126984,
|
| 998 |
+
"eval_logits/chosen": 7.213573455810547,
|
| 999 |
+
"eval_logits/rejected": 7.25667667388916,
|
| 1000 |
+
"eval_logps/chosen": -3.8712658882141113,
|
| 1001 |
+
"eval_logps/rejected": -5.882909774780273,
|
| 1002 |
+
"eval_loss": 0.3922204077243805,
|
| 1003 |
+
"eval_rewards/accuracies": 0.90625,
|
| 1004 |
+
"eval_rewards/chosen": -5.806899070739746,
|
| 1005 |
+
"eval_rewards/margins": 3.017465591430664,
|
| 1006 |
+
"eval_rewards/rejected": -8.824363708496094,
|
| 1007 |
+
"eval_runtime": 18.3231,
|
| 1008 |
+
"eval_samples_per_second": 27.834,
|
| 1009 |
+
"eval_steps_per_second": 3.493,
|
| 1010 |
+
"step": 550
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 0.7111111111111111,
|
| 1014 |
+
"grad_norm": 1.9696826934814453,
|
| 1015 |
+
"learning_rate": 3.4687889661302577e-06,
|
| 1016 |
+
"logits/chosen": 7.829404354095459,
|
| 1017 |
+
"logits/rejected": 7.19503927230835,
|
| 1018 |
+
"logps/chosen": -4.589760780334473,
|
| 1019 |
+
"logps/rejected": -6.637518405914307,
|
| 1020 |
+
"loss": 0.4127,
|
| 1021 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1022 |
+
"rewards/chosen": -6.884641170501709,
|
| 1023 |
+
"rewards/margins": 3.0716373920440674,
|
| 1024 |
+
"rewards/rejected": -9.956277847290039,
|
| 1025 |
+
"step": 560
|
| 1026 |
+
},
|
| 1027 |
+
{
|
| 1028 |
+
"epoch": 0.7238095238095238,
|
| 1029 |
+
"grad_norm": 4.987663269042969,
|
| 1030 |
+
"learning_rate": 3.4203113817116955e-06,
|
| 1031 |
+
"logits/chosen": 7.03420877456665,
|
| 1032 |
+
"logits/rejected": 7.229029655456543,
|
| 1033 |
+
"logps/chosen": -3.7655014991760254,
|
| 1034 |
+
"logps/rejected": -6.0161333084106445,
|
| 1035 |
+
"loss": 0.4436,
|
| 1036 |
+
"rewards/accuracies": 0.9375,
|
| 1037 |
+
"rewards/chosen": -5.648251533508301,
|
| 1038 |
+
"rewards/margins": 3.375948667526245,
|
| 1039 |
+
"rewards/rejected": -9.024200439453125,
|
| 1040 |
+
"step": 570
|
| 1041 |
+
},
|
| 1042 |
+
{
|
| 1043 |
+
"epoch": 0.7365079365079366,
|
| 1044 |
+
"grad_norm": 2.468047618865967,
|
| 1045 |
+
"learning_rate": 3.3714301183045382e-06,
|
| 1046 |
+
"logits/chosen": 6.352839946746826,
|
| 1047 |
+
"logits/rejected": 6.344137668609619,
|
| 1048 |
+
"logps/chosen": -2.9854183197021484,
|
| 1049 |
+
"logps/rejected": -5.105284690856934,
|
| 1050 |
+
"loss": 0.411,
|
| 1051 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 1052 |
+
"rewards/chosen": -4.478127479553223,
|
| 1053 |
+
"rewards/margins": 3.1797995567321777,
|
| 1054 |
+
"rewards/rejected": -7.657926082611084,
|
| 1055 |
+
"step": 580
|
| 1056 |
+
},
|
| 1057 |
+
{
|
| 1058 |
+
"epoch": 0.7492063492063492,
|
| 1059 |
+
"grad_norm": 2.321638822555542,
|
| 1060 |
+
"learning_rate": 3.3221666168464584e-06,
|
| 1061 |
+
"logits/chosen": 7.339761257171631,
|
| 1062 |
+
"logits/rejected": 7.080865383148193,
|
| 1063 |
+
"logps/chosen": -3.7470335960388184,
|
| 1064 |
+
"logps/rejected": -5.2733564376831055,
|
| 1065 |
+
"loss": 0.4297,
|
| 1066 |
+
"rewards/accuracies": 0.8500000238418579,
|
| 1067 |
+
"rewards/chosen": -5.62054967880249,
|
| 1068 |
+
"rewards/margins": 2.2894842624664307,
|
| 1069 |
+
"rewards/rejected": -7.9100341796875,
|
| 1070 |
+
"step": 590
|
| 1071 |
+
},
|
| 1072 |
+
{
|
| 1073 |
+
"epoch": 0.7619047619047619,
|
| 1074 |
+
"grad_norm": 6.850190162658691,
|
| 1075 |
+
"learning_rate": 3.272542485937369e-06,
|
| 1076 |
+
"logits/chosen": 7.278192043304443,
|
| 1077 |
+
"logits/rejected": 6.769750118255615,
|
| 1078 |
+
"logps/chosen": -3.9920971393585205,
|
| 1079 |
+
"logps/rejected": -5.948962211608887,
|
| 1080 |
+
"loss": 0.4537,
|
| 1081 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 1082 |
+
"rewards/chosen": -5.988145351409912,
|
| 1083 |
+
"rewards/margins": 2.9352970123291016,
|
| 1084 |
+
"rewards/rejected": -8.923442840576172,
|
| 1085 |
+
"step": 600
|
| 1086 |
+
},
|
| 1087 |
+
{
|
| 1088 |
+
"epoch": 0.7619047619047619,
|
| 1089 |
+
"eval_logits/chosen": 6.745523929595947,
|
| 1090 |
+
"eval_logits/rejected": 6.732550144195557,
|
| 1091 |
+
"eval_logps/chosen": -3.667492628097534,
|
| 1092 |
+
"eval_logps/rejected": -5.859607696533203,
|
| 1093 |
+
"eval_loss": 0.368425577878952,
|
| 1094 |
+
"eval_rewards/accuracies": 0.921875,
|
| 1095 |
+
"eval_rewards/chosen": -5.501238822937012,
|
| 1096 |
+
"eval_rewards/margins": 3.288173198699951,
|
| 1097 |
+
"eval_rewards/rejected": -8.789411544799805,
|
| 1098 |
+
"eval_runtime": 18.3412,
|
| 1099 |
+
"eval_samples_per_second": 27.806,
|
| 1100 |
+
"eval_steps_per_second": 3.489,
|
| 1101 |
+
"step": 600
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"epoch": 0.7746031746031746,
|
| 1105 |
+
"grad_norm": 3.103388786315918,
|
| 1106 |
+
"learning_rate": 3.222579492361179e-06,
|
| 1107 |
+
"logits/chosen": 6.413697719573975,
|
| 1108 |
+
"logits/rejected": 6.046789646148682,
|
| 1109 |
+
"logps/chosen": -4.207709312438965,
|
| 1110 |
+
"logps/rejected": -6.105495452880859,
|
| 1111 |
+
"loss": 0.4315,
|
| 1112 |
+
"rewards/accuracies": 0.862500011920929,
|
| 1113 |
+
"rewards/chosen": -6.311563968658447,
|
| 1114 |
+
"rewards/margins": 2.846680164337158,
|
| 1115 |
+
"rewards/rejected": -9.158244132995605,
|
| 1116 |
+
"step": 610
|
| 1117 |
+
},
|
| 1118 |
+
{
|
| 1119 |
+
"epoch": 0.7873015873015873,
|
| 1120 |
+
"grad_norm": 4.581955432891846,
|
| 1121 |
+
"learning_rate": 3.1722995515381644e-06,
|
| 1122 |
+
"logits/chosen": 6.37155818939209,
|
| 1123 |
+
"logits/rejected": 6.382074356079102,
|
| 1124 |
+
"logps/chosen": -4.053309917449951,
|
| 1125 |
+
"logps/rejected": -5.929131507873535,
|
| 1126 |
+
"loss": 0.4161,
|
| 1127 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1128 |
+
"rewards/chosen": -6.079965114593506,
|
| 1129 |
+
"rewards/margins": 2.813732624053955,
|
| 1130 |
+
"rewards/rejected": -8.893697738647461,
|
| 1131 |
+
"step": 620
|
| 1132 |
+
},
|
| 1133 |
+
{
|
| 1134 |
+
"epoch": 0.8,
|
| 1135 |
+
"grad_norm": 3.6864757537841797,
|
| 1136 |
+
"learning_rate": 3.121724717912138e-06,
|
| 1137 |
+
"logits/chosen": 5.701592445373535,
|
| 1138 |
+
"logits/rejected": 5.8575944900512695,
|
| 1139 |
+
"logps/chosen": -3.532477617263794,
|
| 1140 |
+
"logps/rejected": -5.849920749664307,
|
| 1141 |
+
"loss": 0.3621,
|
| 1142 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1143 |
+
"rewards/chosen": -5.2987165451049805,
|
| 1144 |
+
"rewards/margins": 3.4761645793914795,
|
| 1145 |
+
"rewards/rejected": -8.774881362915039,
|
| 1146 |
+
"step": 630
|
| 1147 |
+
},
|
| 1148 |
+
{
|
| 1149 |
+
"epoch": 0.8126984126984127,
|
| 1150 |
+
"grad_norm": 3.3179032802581787,
|
| 1151 |
+
"learning_rate": 3.0708771752766397e-06,
|
| 1152 |
+
"logits/chosen": 5.983328819274902,
|
| 1153 |
+
"logits/rejected": 5.853272914886475,
|
| 1154 |
+
"logps/chosen": -3.7127442359924316,
|
| 1155 |
+
"logps/rejected": -5.6633429527282715,
|
| 1156 |
+
"loss": 0.4325,
|
| 1157 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 1158 |
+
"rewards/chosen": -5.569116115570068,
|
| 1159 |
+
"rewards/margins": 2.9258980751037598,
|
| 1160 |
+
"rewards/rejected": -8.495015144348145,
|
| 1161 |
+
"step": 640
|
| 1162 |
+
},
|
| 1163 |
+
{
|
| 1164 |
+
"epoch": 0.8253968253968254,
|
| 1165 |
+
"grad_norm": 3.3010666370391846,
|
| 1166 |
+
"learning_rate": 3.019779227044398e-06,
|
| 1167 |
+
"logits/chosen": 7.427163600921631,
|
| 1168 |
+
"logits/rejected": 7.144589900970459,
|
| 1169 |
+
"logps/chosen": -3.395155429840088,
|
| 1170 |
+
"logps/rejected": -5.485442638397217,
|
| 1171 |
+
"loss": 0.4376,
|
| 1172 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 1173 |
+
"rewards/chosen": -5.0927324295043945,
|
| 1174 |
+
"rewards/margins": 3.1354317665100098,
|
| 1175 |
+
"rewards/rejected": -8.228163719177246,
|
| 1176 |
+
"step": 650
|
| 1177 |
+
},
|
| 1178 |
+
{
|
| 1179 |
+
"epoch": 0.8253968253968254,
|
| 1180 |
+
"eval_logits/chosen": 6.977372169494629,
|
| 1181 |
+
"eval_logits/rejected": 6.941837787628174,
|
| 1182 |
+
"eval_logps/chosen": -3.4915952682495117,
|
| 1183 |
+
"eval_logps/rejected": -5.880876541137695,
|
| 1184 |
+
"eval_loss": 0.3507956266403198,
|
| 1185 |
+
"eval_rewards/accuracies": 0.921875,
|
| 1186 |
+
"eval_rewards/chosen": -5.237393379211426,
|
| 1187 |
+
"eval_rewards/margins": 3.5839221477508545,
|
| 1188 |
+
"eval_rewards/rejected": -8.821314811706543,
|
| 1189 |
+
"eval_runtime": 18.3366,
|
| 1190 |
+
"eval_samples_per_second": 27.813,
|
| 1191 |
+
"eval_steps_per_second": 3.49,
|
| 1192 |
+
"step": 650
|
| 1193 |
+
},
|
| 1194 |
+
{
|
| 1195 |
+
"epoch": 0.8380952380952381,
|
| 1196 |
+
"grad_norm": 6.439872741699219,
|
| 1197 |
+
"learning_rate": 2.9684532864643123e-06,
|
| 1198 |
+
"logits/chosen": 7.413880825042725,
|
| 1199 |
+
"logits/rejected": 7.040737152099609,
|
| 1200 |
+
"logps/chosen": -4.771336078643799,
|
| 1201 |
+
"logps/rejected": -7.235314846038818,
|
| 1202 |
+
"loss": 0.381,
|
| 1203 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 1204 |
+
"rewards/chosen": -7.157004356384277,
|
| 1205 |
+
"rewards/margins": 3.6959691047668457,
|
| 1206 |
+
"rewards/rejected": -10.852972984313965,
|
| 1207 |
+
"step": 660
|
| 1208 |
+
},
|
| 1209 |
+
{
|
| 1210 |
+
"epoch": 0.8507936507936508,
|
| 1211 |
+
"grad_norm": 5.114463806152344,
|
| 1212 |
+
"learning_rate": 2.9169218667902562e-06,
|
| 1213 |
+
"logits/chosen": 6.7717437744140625,
|
| 1214 |
+
"logits/rejected": 6.568068504333496,
|
| 1215 |
+
"logps/chosen": -3.372582197189331,
|
| 1216 |
+
"logps/rejected": -5.318942070007324,
|
| 1217 |
+
"loss": 0.4385,
|
| 1218 |
+
"rewards/accuracies": 0.8500000238418579,
|
| 1219 |
+
"rewards/chosen": -5.058873653411865,
|
| 1220 |
+
"rewards/margins": 2.9195384979248047,
|
| 1221 |
+
"rewards/rejected": -7.9784135818481445,
|
| 1222 |
+
"step": 670
|
| 1223 |
+
},
|
| 1224 |
+
{
|
| 1225 |
+
"epoch": 0.8634920634920635,
|
| 1226 |
+
"grad_norm": 2.7731258869171143,
|
| 1227 |
+
"learning_rate": 2.8652075714060296e-06,
|
| 1228 |
+
"logits/chosen": 6.759649753570557,
|
| 1229 |
+
"logits/rejected": 6.250086784362793,
|
| 1230 |
+
"logps/chosen": -3.9128894805908203,
|
| 1231 |
+
"logps/rejected": -6.746951103210449,
|
| 1232 |
+
"loss": 0.3879,
|
| 1233 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1234 |
+
"rewards/chosen": -5.8693342208862305,
|
| 1235 |
+
"rewards/margins": 4.251092433929443,
|
| 1236 |
+
"rewards/rejected": -10.120426177978516,
|
| 1237 |
+
"step": 680
|
| 1238 |
+
},
|
| 1239 |
+
{
|
| 1240 |
+
"epoch": 0.8761904761904762,
|
| 1241 |
+
"grad_norm": 2.517974615097046,
|
| 1242 |
+
"learning_rate": 2.813333083910761e-06,
|
| 1243 |
+
"logits/chosen": 7.3841047286987305,
|
| 1244 |
+
"logits/rejected": 7.0824761390686035,
|
| 1245 |
+
"logps/chosen": -4.643451690673828,
|
| 1246 |
+
"logps/rejected": -6.954865455627441,
|
| 1247 |
+
"loss": 0.3451,
|
| 1248 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1249 |
+
"rewards/chosen": -6.965176582336426,
|
| 1250 |
+
"rewards/margins": 3.467121124267578,
|
| 1251 |
+
"rewards/rejected": -10.43229866027832,
|
| 1252 |
+
"step": 690
|
| 1253 |
+
},
|
| 1254 |
+
{
|
| 1255 |
+
"epoch": 0.8888888888888888,
|
| 1256 |
+
"grad_norm": 4.514930725097656,
|
| 1257 |
+
"learning_rate": 2.761321158169134e-06,
|
| 1258 |
+
"logits/chosen": 6.828726291656494,
|
| 1259 |
+
"logits/rejected": 6.3069562911987305,
|
| 1260 |
+
"logps/chosen": -4.702274322509766,
|
| 1261 |
+
"logps/rejected": -7.075028896331787,
|
| 1262 |
+
"loss": 0.4094,
|
| 1263 |
+
"rewards/accuracies": 0.8500000238418579,
|
| 1264 |
+
"rewards/chosen": -7.053411960601807,
|
| 1265 |
+
"rewards/margins": 3.5591323375701904,
|
| 1266 |
+
"rewards/rejected": -10.612543106079102,
|
| 1267 |
+
"step": 700
|
| 1268 |
+
},
|
| 1269 |
+
{
|
| 1270 |
+
"epoch": 0.8888888888888888,
|
| 1271 |
+
"eval_logits/chosen": 6.324803352355957,
|
| 1272 |
+
"eval_logits/rejected": 6.258608818054199,
|
| 1273 |
+
"eval_logps/chosen": -3.811472177505493,
|
| 1274 |
+
"eval_logps/rejected": -6.270685195922852,
|
| 1275 |
+
"eval_loss": 0.33448922634124756,
|
| 1276 |
+
"eval_rewards/accuracies": 0.9375,
|
| 1277 |
+
"eval_rewards/chosen": -5.717207431793213,
|
| 1278 |
+
"eval_rewards/margins": 3.6888198852539062,
|
| 1279 |
+
"eval_rewards/rejected": -9.406026840209961,
|
| 1280 |
+
"eval_runtime": 18.3448,
|
| 1281 |
+
"eval_samples_per_second": 27.801,
|
| 1282 |
+
"eval_steps_per_second": 3.489,
|
| 1283 |
+
"step": 700
|
| 1284 |
+
},
|
| 1285 |
+
{
|
| 1286 |
+
"epoch": 0.9015873015873016,
|
| 1287 |
+
"grad_norm": 2.871471405029297,
|
| 1288 |
+
"learning_rate": 2.70919460833079e-06,
|
| 1289 |
+
"logits/chosen": 6.418933868408203,
|
| 1290 |
+
"logits/rejected": 6.141990661621094,
|
| 1291 |
+
"logps/chosen": -4.123613357543945,
|
| 1292 |
+
"logps/rejected": -6.900608062744141,
|
| 1293 |
+
"loss": 0.3604,
|
| 1294 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1295 |
+
"rewards/chosen": -6.18541955947876,
|
| 1296 |
+
"rewards/margins": 4.165491580963135,
|
| 1297 |
+
"rewards/rejected": -10.350912094116211,
|
| 1298 |
+
"step": 710
|
| 1299 |
+
},
|
| 1300 |
+
{
|
| 1301 |
+
"epoch": 0.9142857142857143,
|
| 1302 |
+
"grad_norm": 5.5512824058532715,
|
| 1303 |
+
"learning_rate": 2.6569762988232838e-06,
|
| 1304 |
+
"logits/chosen": 6.1132917404174805,
|
| 1305 |
+
"logits/rejected": 6.01393461227417,
|
| 1306 |
+
"logps/chosen": -4.065993309020996,
|
| 1307 |
+
"logps/rejected": -6.6075005531311035,
|
| 1308 |
+
"loss": 0.3682,
|
| 1309 |
+
"rewards/accuracies": 0.9375,
|
| 1310 |
+
"rewards/chosen": -6.098989963531494,
|
| 1311 |
+
"rewards/margins": 3.812260150909424,
|
| 1312 |
+
"rewards/rejected": -9.911249160766602,
|
| 1313 |
+
"step": 720
|
| 1314 |
+
},
|
| 1315 |
+
{
|
| 1316 |
+
"epoch": 0.926984126984127,
|
| 1317 |
+
"grad_norm": 2.5356554985046387,
|
| 1318 |
+
"learning_rate": 2.604689134322999e-06,
|
| 1319 |
+
"logits/chosen": 6.585450649261475,
|
| 1320 |
+
"logits/rejected": 5.800690650939941,
|
| 1321 |
+
"logps/chosen": -4.80672550201416,
|
| 1322 |
+
"logps/rejected": -6.856999397277832,
|
| 1323 |
+
"loss": 0.3886,
|
| 1324 |
+
"rewards/accuracies": 0.8500000238418579,
|
| 1325 |
+
"rewards/chosen": -7.210087776184082,
|
| 1326 |
+
"rewards/margins": 3.075411319732666,
|
| 1327 |
+
"rewards/rejected": -10.28549861907959,
|
| 1328 |
+
"step": 730
|
| 1329 |
+
},
|
| 1330 |
+
{
|
| 1331 |
+
"epoch": 0.9396825396825397,
|
| 1332 |
+
"grad_norm": 3.685417413711548,
|
| 1333 |
+
"learning_rate": 2.5523560497083927e-06,
|
| 1334 |
+
"logits/chosen": 5.728347301483154,
|
| 1335 |
+
"logits/rejected": 5.568971157073975,
|
| 1336 |
+
"logps/chosen": -4.184335231781006,
|
| 1337 |
+
"logps/rejected": -6.79797887802124,
|
| 1338 |
+
"loss": 0.3548,
|
| 1339 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 1340 |
+
"rewards/chosen": -6.27650260925293,
|
| 1341 |
+
"rewards/margins": 3.9204659461975098,
|
| 1342 |
+
"rewards/rejected": -10.196969032287598,
|
| 1343 |
+
"step": 740
|
| 1344 |
+
},
|
| 1345 |
+
{
|
| 1346 |
+
"epoch": 0.9523809523809523,
|
| 1347 |
+
"grad_norm": 3.107419490814209,
|
| 1348 |
+
"learning_rate": 2.5e-06,
|
| 1349 |
+
"logits/chosen": 6.063660621643066,
|
| 1350 |
+
"logits/rejected": 5.694943428039551,
|
| 1351 |
+
"logps/chosen": -3.863433837890625,
|
| 1352 |
+
"logps/rejected": -6.4726691246032715,
|
| 1353 |
+
"loss": 0.3706,
|
| 1354 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1355 |
+
"rewards/chosen": -5.7951507568359375,
|
| 1356 |
+
"rewards/margins": 3.9138541221618652,
|
| 1357 |
+
"rewards/rejected": -9.709004402160645,
|
| 1358 |
+
"step": 750
|
| 1359 |
+
},
|
| 1360 |
+
{
|
| 1361 |
+
"epoch": 0.9523809523809523,
|
| 1362 |
+
"eval_logits/chosen": 6.257328033447266,
|
| 1363 |
+
"eval_logits/rejected": 6.1753411293029785,
|
| 1364 |
+
"eval_logps/chosen": -3.6567726135253906,
|
| 1365 |
+
"eval_logps/rejected": -6.332419395446777,
|
| 1366 |
+
"eval_loss": 0.31834185123443604,
|
| 1367 |
+
"eval_rewards/accuracies": 0.9375,
|
| 1368 |
+
"eval_rewards/chosen": -5.485158920288086,
|
| 1369 |
+
"eval_rewards/margins": 4.013469696044922,
|
| 1370 |
+
"eval_rewards/rejected": -9.498628616333008,
|
| 1371 |
+
"eval_runtime": 18.3371,
|
| 1372 |
+
"eval_samples_per_second": 27.812,
|
| 1373 |
+
"eval_steps_per_second": 3.49,
|
| 1374 |
+
"step": 750
|
| 1375 |
+
},
|
| 1376 |
+
{
|
| 1377 |
+
"epoch": 0.9650793650793651,
|
| 1378 |
+
"grad_norm": 5.022021293640137,
|
| 1379 |
+
"learning_rate": 2.447643950291608e-06,
|
| 1380 |
+
"logits/chosen": 6.415736198425293,
|
| 1381 |
+
"logits/rejected": 6.128665924072266,
|
| 1382 |
+
"logps/chosen": -3.9557278156280518,
|
| 1383 |
+
"logps/rejected": -6.381206035614014,
|
| 1384 |
+
"loss": 0.3938,
|
| 1385 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 1386 |
+
"rewards/chosen": -5.933591365814209,
|
| 1387 |
+
"rewards/margins": 3.6382179260253906,
|
| 1388 |
+
"rewards/rejected": -9.571809768676758,
|
| 1389 |
+
"step": 760
|
| 1390 |
+
},
|
| 1391 |
+
{
|
| 1392 |
+
"epoch": 0.9777777777777777,
|
| 1393 |
+
"grad_norm": 4.283695220947266,
|
| 1394 |
+
"learning_rate": 2.3953108656770018e-06,
|
| 1395 |
+
"logits/chosen": 5.191827774047852,
|
| 1396 |
+
"logits/rejected": 5.130216598510742,
|
| 1397 |
+
"logps/chosen": -3.842036724090576,
|
| 1398 |
+
"logps/rejected": -6.3525071144104,
|
| 1399 |
+
"loss": 0.3744,
|
| 1400 |
+
"rewards/accuracies": 0.887499988079071,
|
| 1401 |
+
"rewards/chosen": -5.763054847717285,
|
| 1402 |
+
"rewards/margins": 3.7657058238983154,
|
| 1403 |
+
"rewards/rejected": -9.52876091003418,
|
| 1404 |
+
"step": 770
|
| 1405 |
+
},
|
| 1406 |
+
{
|
| 1407 |
+
"epoch": 0.9904761904761905,
|
| 1408 |
+
"grad_norm": 4.412943363189697,
|
| 1409 |
+
"learning_rate": 2.3430237011767166e-06,
|
| 1410 |
+
"logits/chosen": 6.2709760665893555,
|
| 1411 |
+
"logits/rejected": 5.860433101654053,
|
| 1412 |
+
"logps/chosen": -4.709076881408691,
|
| 1413 |
+
"logps/rejected": -7.461434841156006,
|
| 1414 |
+
"loss": 0.3914,
|
| 1415 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1416 |
+
"rewards/chosen": -7.063615322113037,
|
| 1417 |
+
"rewards/margins": 4.128536224365234,
|
| 1418 |
+
"rewards/rejected": -11.19215202331543,
|
| 1419 |
+
"step": 780
|
| 1420 |
+
},
|
| 1421 |
+
{
|
| 1422 |
+
"epoch": 1.0025396825396826,
|
| 1423 |
+
"grad_norm": 2.1262686252593994,
|
| 1424 |
+
"learning_rate": 2.290805391669212e-06,
|
| 1425 |
+
"logits/chosen": 6.877432346343994,
|
| 1426 |
+
"logits/rejected": 6.122270584106445,
|
| 1427 |
+
"logps/chosen": -4.66621732711792,
|
| 1428 |
+
"logps/rejected": -7.71566915512085,
|
| 1429 |
+
"loss": 0.3467,
|
| 1430 |
+
"rewards/accuracies": 0.8947368264198303,
|
| 1431 |
+
"rewards/chosen": -6.999325275421143,
|
| 1432 |
+
"rewards/margins": 4.574176788330078,
|
| 1433 |
+
"rewards/rejected": -11.573503494262695,
|
| 1434 |
+
"step": 790
|
| 1435 |
+
},
|
| 1436 |
+
{
|
| 1437 |
+
"epoch": 1.0152380952380953,
|
| 1438 |
+
"grad_norm": 4.042492389678955,
|
| 1439 |
+
"learning_rate": 2.238678841830867e-06,
|
| 1440 |
+
"logits/chosen": 6.55672550201416,
|
| 1441 |
+
"logits/rejected": 5.752241134643555,
|
| 1442 |
+
"logps/chosen": -4.66592264175415,
|
| 1443 |
+
"logps/rejected": -6.966067314147949,
|
| 1444 |
+
"loss": 0.3404,
|
| 1445 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 1446 |
+
"rewards/chosen": -6.998884677886963,
|
| 1447 |
+
"rewards/margins": 3.4502170085906982,
|
| 1448 |
+
"rewards/rejected": -10.449102401733398,
|
| 1449 |
+
"step": 800
|
| 1450 |
+
},
|
| 1451 |
+
{
|
| 1452 |
+
"epoch": 1.0152380952380953,
|
| 1453 |
+
"eval_logits/chosen": 6.143824100494385,
|
| 1454 |
+
"eval_logits/rejected": 6.0408477783203125,
|
| 1455 |
+
"eval_logps/chosen": -3.900071144104004,
|
| 1456 |
+
"eval_logps/rejected": -6.64994478225708,
|
| 1457 |
+
"eval_loss": 0.30234017968177795,
|
| 1458 |
+
"eval_rewards/accuracies": 0.9375,
|
| 1459 |
+
"eval_rewards/chosen": -5.850106239318848,
|
| 1460 |
+
"eval_rewards/margins": 4.124810695648193,
|
| 1461 |
+
"eval_rewards/rejected": -9.974918365478516,
|
| 1462 |
+
"eval_runtime": 18.317,
|
| 1463 |
+
"eval_samples_per_second": 27.843,
|
| 1464 |
+
"eval_steps_per_second": 3.494,
|
| 1465 |
+
"step": 800
|
| 1466 |
+
},
|
| 1467 |
+
{
|
| 1468 |
+
"epoch": 1.027936507936508,
|
| 1469 |
+
"grad_norm": 4.464768886566162,
|
| 1470 |
+
"learning_rate": 2.186666916089239e-06,
|
| 1471 |
+
"logits/chosen": 5.800255298614502,
|
| 1472 |
+
"logits/rejected": 5.524979114532471,
|
| 1473 |
+
"logps/chosen": -4.174003601074219,
|
| 1474 |
+
"logps/rejected": -6.4187140464782715,
|
| 1475 |
+
"loss": 0.3517,
|
| 1476 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 1477 |
+
"rewards/chosen": -6.261005401611328,
|
| 1478 |
+
"rewards/margins": 3.367064952850342,
|
| 1479 |
+
"rewards/rejected": -9.628071784973145,
|
| 1480 |
+
"step": 810
|
| 1481 |
+
},
|
| 1482 |
+
{
|
| 1483 |
+
"epoch": 1.0406349206349206,
|
| 1484 |
+
"grad_norm": 2.864623546600342,
|
| 1485 |
+
"learning_rate": 2.134792428593971e-06,
|
| 1486 |
+
"logits/chosen": 5.549002170562744,
|
| 1487 |
+
"logits/rejected": 5.762969017028809,
|
| 1488 |
+
"logps/chosen": -4.055948734283447,
|
| 1489 |
+
"logps/rejected": -6.332015037536621,
|
| 1490 |
+
"loss": 0.3525,
|
| 1491 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 1492 |
+
"rewards/chosen": -6.083922386169434,
|
| 1493 |
+
"rewards/margins": 3.4140994548797607,
|
| 1494 |
+
"rewards/rejected": -9.498022079467773,
|
| 1495 |
+
"step": 820
|
| 1496 |
+
},
|
| 1497 |
+
{
|
| 1498 |
+
"epoch": 1.0533333333333332,
|
| 1499 |
+
"grad_norm": 4.997330188751221,
|
| 1500 |
+
"learning_rate": 2.0830781332097446e-06,
|
| 1501 |
+
"logits/chosen": 5.439722061157227,
|
| 1502 |
+
"logits/rejected": 5.755289554595947,
|
| 1503 |
+
"logps/chosen": -4.239761829376221,
|
| 1504 |
+
"logps/rejected": -6.699388027191162,
|
| 1505 |
+
"loss": 0.3423,
|
| 1506 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 1507 |
+
"rewards/chosen": -6.359642028808594,
|
| 1508 |
+
"rewards/margins": 3.689438581466675,
|
| 1509 |
+
"rewards/rejected": -10.049081802368164,
|
| 1510 |
+
"step": 830
|
| 1511 |
+
},
|
| 1512 |
+
{
|
| 1513 |
+
"epoch": 1.066031746031746,
|
| 1514 |
+
"grad_norm": 4.705864906311035,
|
| 1515 |
+
"learning_rate": 2.031546713535688e-06,
|
| 1516 |
+
"logits/chosen": 6.309781074523926,
|
| 1517 |
+
"logits/rejected": 6.002473831176758,
|
| 1518 |
+
"logps/chosen": -3.9887123107910156,
|
| 1519 |
+
"logps/rejected": -6.414917945861816,
|
| 1520 |
+
"loss": 0.327,
|
| 1521 |
+
"rewards/accuracies": 0.9375,
|
| 1522 |
+
"rewards/chosen": -5.983067989349365,
|
| 1523 |
+
"rewards/margins": 3.639307737350464,
|
| 1524 |
+
"rewards/rejected": -9.622376441955566,
|
| 1525 |
+
"step": 840
|
| 1526 |
+
},
|
| 1527 |
+
{
|
| 1528 |
+
"epoch": 1.0787301587301588,
|
| 1529 |
+
"grad_norm": 4.2711286544799805,
|
| 1530 |
+
"learning_rate": 1.9802207729556023e-06,
|
| 1531 |
+
"logits/chosen": 6.275202751159668,
|
| 1532 |
+
"logits/rejected": 5.929355621337891,
|
| 1533 |
+
"logps/chosen": -4.061328887939453,
|
| 1534 |
+
"logps/rejected": -6.566980838775635,
|
| 1535 |
+
"loss": 0.3587,
|
| 1536 |
+
"rewards/accuracies": 0.875,
|
| 1537 |
+
"rewards/chosen": -6.09199333190918,
|
| 1538 |
+
"rewards/margins": 3.758476734161377,
|
| 1539 |
+
"rewards/rejected": -9.850470542907715,
|
| 1540 |
+
"step": 850
|
| 1541 |
+
},
|
| 1542 |
+
{
|
| 1543 |
+
"epoch": 1.0787301587301588,
|
| 1544 |
+
"eval_logits/chosen": 5.9675092697143555,
|
| 1545 |
+
"eval_logits/rejected": 5.841681480407715,
|
| 1546 |
+
"eval_logps/chosen": -3.5920660495758057,
|
| 1547 |
+
"eval_logps/rejected": -6.45810604095459,
|
| 1548 |
+
"eval_loss": 0.2958359122276306,
|
| 1549 |
+
"eval_rewards/accuracies": 0.9375,
|
| 1550 |
+
"eval_rewards/chosen": -5.38809871673584,
|
| 1551 |
+
"eval_rewards/margins": 4.2990593910217285,
|
| 1552 |
+
"eval_rewards/rejected": -9.687158584594727,
|
| 1553 |
+
"eval_runtime": 18.3401,
|
| 1554 |
+
"eval_samples_per_second": 27.808,
|
| 1555 |
+
"eval_steps_per_second": 3.49,
|
| 1556 |
+
"step": 850
|
| 1557 |
+
},
|
| 1558 |
+
{
|
| 1559 |
+
"epoch": 1.0914285714285714,
|
| 1560 |
+
"grad_norm": 2.666412115097046,
|
| 1561 |
+
"learning_rate": 1.9291228247233607e-06,
|
| 1562 |
+
"logits/chosen": 6.118748188018799,
|
| 1563 |
+
"logits/rejected": 5.885035991668701,
|
| 1564 |
+
"logps/chosen": -3.863534927368164,
|
| 1565 |
+
"logps/rejected": -6.175184726715088,
|
| 1566 |
+
"loss": 0.3297,
|
| 1567 |
+
"rewards/accuracies": 0.875,
|
| 1568 |
+
"rewards/chosen": -5.795302391052246,
|
| 1569 |
+
"rewards/margins": 3.467475175857544,
|
| 1570 |
+
"rewards/rejected": -9.262777328491211,
|
| 1571 |
+
"step": 860
|
| 1572 |
+
},
|
| 1573 |
+
{
|
| 1574 |
+
"epoch": 1.104126984126984,
|
| 1575 |
+
"grad_norm": 2.253833293914795,
|
| 1576 |
+
"learning_rate": 1.8782752820878636e-06,
|
| 1577 |
+
"logits/chosen": 6.99627685546875,
|
| 1578 |
+
"logits/rejected": 6.862641334533691,
|
| 1579 |
+
"logps/chosen": -3.3962769508361816,
|
| 1580 |
+
"logps/rejected": -5.6661272048950195,
|
| 1581 |
+
"loss": 0.339,
|
| 1582 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 1583 |
+
"rewards/chosen": -5.094415187835693,
|
| 1584 |
+
"rewards/margins": 3.404775619506836,
|
| 1585 |
+
"rewards/rejected": -8.499190330505371,
|
| 1586 |
+
"step": 870
|
| 1587 |
+
},
|
| 1588 |
+
{
|
| 1589 |
+
"epoch": 1.116825396825397,
|
| 1590 |
+
"grad_norm": 4.063273906707764,
|
| 1591 |
+
"learning_rate": 1.827700448461836e-06,
|
| 1592 |
+
"logits/chosen": 5.741837024688721,
|
| 1593 |
+
"logits/rejected": 5.7893548011779785,
|
| 1594 |
+
"logps/chosen": -3.724289655685425,
|
| 1595 |
+
"logps/rejected": -6.367809772491455,
|
| 1596 |
+
"loss": 0.304,
|
| 1597 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1598 |
+
"rewards/chosen": -5.586434841156006,
|
| 1599 |
+
"rewards/margins": 3.965280532836914,
|
| 1600 |
+
"rewards/rejected": -9.551714897155762,
|
| 1601 |
+
"step": 880
|
| 1602 |
+
},
|
| 1603 |
+
{
|
| 1604 |
+
"epoch": 1.1295238095238096,
|
| 1605 |
+
"grad_norm": 4.975926399230957,
|
| 1606 |
+
"learning_rate": 1.7774205076388207e-06,
|
| 1607 |
+
"logits/chosen": 6.33252477645874,
|
| 1608 |
+
"logits/rejected": 6.018566131591797,
|
| 1609 |
+
"logps/chosen": -4.238375186920166,
|
| 1610 |
+
"logps/rejected": -7.243282318115234,
|
| 1611 |
+
"loss": 0.3124,
|
| 1612 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 1613 |
+
"rewards/chosen": -6.35756254196167,
|
| 1614 |
+
"rewards/margins": 4.50736141204834,
|
| 1615 |
+
"rewards/rejected": -10.864924430847168,
|
| 1616 |
+
"step": 890
|
| 1617 |
+
},
|
| 1618 |
+
{
|
| 1619 |
+
"epoch": 1.1422222222222222,
|
| 1620 |
+
"grad_norm": 3.159183979034424,
|
| 1621 |
+
"learning_rate": 1.7274575140626318e-06,
|
| 1622 |
+
"logits/chosen": 6.32235860824585,
|
| 1623 |
+
"logits/rejected": 5.313261985778809,
|
| 1624 |
+
"logps/chosen": -4.42888879776001,
|
| 1625 |
+
"logps/rejected": -7.219472408294678,
|
| 1626 |
+
"loss": 0.3122,
|
| 1627 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 1628 |
+
"rewards/chosen": -6.643332481384277,
|
| 1629 |
+
"rewards/margins": 4.185876846313477,
|
| 1630 |
+
"rewards/rejected": -10.82921028137207,
|
| 1631 |
+
"step": 900
|
| 1632 |
+
},
|
| 1633 |
+
{
|
| 1634 |
+
"epoch": 1.1422222222222222,
|
| 1635 |
+
"eval_logits/chosen": 5.802257537841797,
|
| 1636 |
+
"eval_logits/rejected": 5.6801910400390625,
|
| 1637 |
+
"eval_logps/chosen": -3.7256534099578857,
|
| 1638 |
+
"eval_logps/rejected": -6.689548492431641,
|
| 1639 |
+
"eval_loss": 0.28521302342414856,
|
| 1640 |
+
"eval_rewards/accuracies": 0.9375,
|
| 1641 |
+
"eval_rewards/chosen": -5.588479995727539,
|
| 1642 |
+
"eval_rewards/margins": 4.445842742919922,
|
| 1643 |
+
"eval_rewards/rejected": -10.034323692321777,
|
| 1644 |
+
"eval_runtime": 18.2302,
|
| 1645 |
+
"eval_samples_per_second": 27.976,
|
| 1646 |
+
"eval_steps_per_second": 3.511,
|
| 1647 |
+
"step": 900
|
| 1648 |
+
},
|
| 1649 |
+
{
|
| 1650 |
+
"epoch": 1.154920634920635,
|
| 1651 |
+
"grad_norm": 5.580939292907715,
|
| 1652 |
+
"learning_rate": 1.677833383153542e-06,
|
| 1653 |
+
"logits/chosen": 6.4255805015563965,
|
| 1654 |
+
"logits/rejected": 5.784458160400391,
|
| 1655 |
+
"logps/chosen": -4.204034328460693,
|
| 1656 |
+
"logps/rejected": -6.822949409484863,
|
| 1657 |
+
"loss": 0.3277,
|
| 1658 |
+
"rewards/accuracies": 0.875,
|
| 1659 |
+
"rewards/chosen": -6.306051731109619,
|
| 1660 |
+
"rewards/margins": 3.928372621536255,
|
| 1661 |
+
"rewards/rejected": -10.234424591064453,
|
| 1662 |
+
"step": 910
|
| 1663 |
+
},
|
| 1664 |
+
{
|
| 1665 |
+
"epoch": 1.1676190476190476,
|
| 1666 |
+
"grad_norm": 3.5059311389923096,
|
| 1667 |
+
"learning_rate": 1.6285698816954626e-06,
|
| 1668 |
+
"logits/chosen": 5.253472328186035,
|
| 1669 |
+
"logits/rejected": 5.226204872131348,
|
| 1670 |
+
"logps/chosen": -3.458686113357544,
|
| 1671 |
+
"logps/rejected": -5.801349639892578,
|
| 1672 |
+
"loss": 0.3212,
|
| 1673 |
+
"rewards/accuracies": 0.9375,
|
| 1674 |
+
"rewards/chosen": -5.188028812408447,
|
| 1675 |
+
"rewards/margins": 3.51399564743042,
|
| 1676 |
+
"rewards/rejected": -8.702024459838867,
|
| 1677 |
+
"step": 920
|
| 1678 |
+
},
|
| 1679 |
+
{
|
| 1680 |
+
"epoch": 1.1803174603174602,
|
| 1681 |
+
"grad_norm": 5.19541072845459,
|
| 1682 |
+
"learning_rate": 1.5796886182883053e-06,
|
| 1683 |
+
"logits/chosen": 5.724546909332275,
|
| 1684 |
+
"logits/rejected": 5.355966091156006,
|
| 1685 |
+
"logps/chosen": -4.142882347106934,
|
| 1686 |
+
"logps/rejected": -7.1851396560668945,
|
| 1687 |
+
"loss": 0.3472,
|
| 1688 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1689 |
+
"rewards/chosen": -6.214322566986084,
|
| 1690 |
+
"rewards/margins": 4.563385963439941,
|
| 1691 |
+
"rewards/rejected": -10.777708053588867,
|
| 1692 |
+
"step": 930
|
| 1693 |
+
},
|
| 1694 |
+
{
|
| 1695 |
+
"epoch": 1.193015873015873,
|
| 1696 |
+
"grad_norm": 3.7898037433624268,
|
| 1697 |
+
"learning_rate": 1.5312110338697427e-06,
|
| 1698 |
+
"logits/chosen": 5.5762104988098145,
|
| 1699 |
+
"logits/rejected": 5.337274551391602,
|
| 1700 |
+
"logps/chosen": -3.7495741844177246,
|
| 1701 |
+
"logps/rejected": -6.201944351196289,
|
| 1702 |
+
"loss": 0.3193,
|
| 1703 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 1704 |
+
"rewards/chosen": -5.624361515045166,
|
| 1705 |
+
"rewards/margins": 3.6785552501678467,
|
| 1706 |
+
"rewards/rejected": -9.30291748046875,
|
| 1707 |
+
"step": 940
|
| 1708 |
+
},
|
| 1709 |
+
{
|
| 1710 |
+
"epoch": 1.2057142857142857,
|
| 1711 |
+
"grad_norm": 8.723204612731934,
|
| 1712 |
+
"learning_rate": 1.4831583923105e-06,
|
| 1713 |
+
"logits/chosen": 6.4793524742126465,
|
| 1714 |
+
"logits/rejected": 6.682563781738281,
|
| 1715 |
+
"logps/chosen": -3.6695919036865234,
|
| 1716 |
+
"logps/rejected": -6.200020790100098,
|
| 1717 |
+
"loss": 0.3267,
|
| 1718 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 1719 |
+
"rewards/chosen": -5.504387855529785,
|
| 1720 |
+
"rewards/margins": 3.795642375946045,
|
| 1721 |
+
"rewards/rejected": -9.300030708312988,
|
| 1722 |
+
"step": 950
|
| 1723 |
+
},
|
| 1724 |
+
{
|
| 1725 |
+
"epoch": 1.2057142857142857,
|
| 1726 |
+
"eval_logits/chosen": 5.892926216125488,
|
| 1727 |
+
"eval_logits/rejected": 5.720220565795898,
|
| 1728 |
+
"eval_logps/chosen": -3.7383785247802734,
|
| 1729 |
+
"eval_logps/rejected": -6.7838873863220215,
|
| 1730 |
+
"eval_loss": 0.27200350165367126,
|
| 1731 |
+
"eval_rewards/accuracies": 0.9375,
|
| 1732 |
+
"eval_rewards/chosen": -5.60756778717041,
|
| 1733 |
+
"eval_rewards/margins": 4.568263053894043,
|
| 1734 |
+
"eval_rewards/rejected": -10.175830841064453,
|
| 1735 |
+
"eval_runtime": 18.3385,
|
| 1736 |
+
"eval_samples_per_second": 27.81,
|
| 1737 |
+
"eval_steps_per_second": 3.49,
|
| 1738 |
+
"step": 950
|
| 1739 |
+
},
|
| 1740 |
+
{
|
| 1741 |
+
"epoch": 1.2184126984126984,
|
| 1742 |
+
"grad_norm": 4.321587085723877,
|
| 1743 |
+
"learning_rate": 1.4355517710873184e-06,
|
| 1744 |
+
"logits/chosen": 5.534070014953613,
|
| 1745 |
+
"logits/rejected": 5.096575736999512,
|
| 1746 |
+
"logps/chosen": -4.546253204345703,
|
| 1747 |
+
"logps/rejected": -7.369414329528809,
|
| 1748 |
+
"loss": 0.3554,
|
| 1749 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1750 |
+
"rewards/chosen": -6.8193793296813965,
|
| 1751 |
+
"rewards/margins": 4.2347412109375,
|
| 1752 |
+
"rewards/rejected": -11.054121017456055,
|
| 1753 |
+
"step": 960
|
| 1754 |
+
},
|
| 1755 |
+
{
|
| 1756 |
+
"epoch": 1.231111111111111,
|
| 1757 |
+
"grad_norm": 2.3186330795288086,
|
| 1758 |
+
"learning_rate": 1.388412052037682e-06,
|
| 1759 |
+
"logits/chosen": 6.364870548248291,
|
| 1760 |
+
"logits/rejected": 6.031035423278809,
|
| 1761 |
+
"logps/chosen": -3.9153428077697754,
|
| 1762 |
+
"logps/rejected": -6.4653639793396,
|
| 1763 |
+
"loss": 0.2867,
|
| 1764 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1765 |
+
"rewards/chosen": -5.873013973236084,
|
| 1766 |
+
"rewards/margins": 3.825031280517578,
|
| 1767 |
+
"rewards/rejected": -9.69804573059082,
|
| 1768 |
+
"step": 970
|
| 1769 |
+
},
|
| 1770 |
+
{
|
| 1771 |
+
"epoch": 1.243809523809524,
|
| 1772 |
+
"grad_norm": 6.711607456207275,
|
| 1773 |
+
"learning_rate": 1.3417599122003464e-06,
|
| 1774 |
+
"logits/chosen": 6.216970443725586,
|
| 1775 |
+
"logits/rejected": 5.668035507202148,
|
| 1776 |
+
"logps/chosen": -4.147327899932861,
|
| 1777 |
+
"logps/rejected": -7.050875186920166,
|
| 1778 |
+
"loss": 0.3064,
|
| 1779 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1780 |
+
"rewards/chosen": -6.220992088317871,
|
| 1781 |
+
"rewards/margins": 4.355320930480957,
|
| 1782 |
+
"rewards/rejected": -10.576312065124512,
|
| 1783 |
+
"step": 980
|
| 1784 |
+
},
|
| 1785 |
+
{
|
| 1786 |
+
"epoch": 1.2565079365079366,
|
| 1787 |
+
"grad_norm": 4.946932792663574,
|
| 1788 |
+
"learning_rate": 1.2956158147457116e-06,
|
| 1789 |
+
"logits/chosen": 5.909533500671387,
|
| 1790 |
+
"logits/rejected": 5.650919437408447,
|
| 1791 |
+
"logps/chosen": -4.229696273803711,
|
| 1792 |
+
"logps/rejected": -7.213942050933838,
|
| 1793 |
+
"loss": 0.3079,
|
| 1794 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 1795 |
+
"rewards/chosen": -6.344544410705566,
|
| 1796 |
+
"rewards/margins": 4.476368427276611,
|
| 1797 |
+
"rewards/rejected": -10.820913314819336,
|
| 1798 |
+
"step": 990
|
| 1799 |
+
},
|
| 1800 |
+
{
|
| 1801 |
+
"epoch": 1.2692063492063492,
|
| 1802 |
+
"grad_norm": 3.831132411956787,
|
| 1803 |
+
"learning_rate": 1.2500000000000007e-06,
|
| 1804 |
+
"logits/chosen": 4.852212429046631,
|
| 1805 |
+
"logits/rejected": 4.722367763519287,
|
| 1806 |
+
"logps/chosen": -4.436448574066162,
|
| 1807 |
+
"logps/rejected": -7.6168317794799805,
|
| 1808 |
+
"loss": 0.3158,
|
| 1809 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 1810 |
+
"rewards/chosen": -6.6546735763549805,
|
| 1811 |
+
"rewards/margins": 4.770575046539307,
|
| 1812 |
+
"rewards/rejected": -11.425248146057129,
|
| 1813 |
+
"step": 1000
|
| 1814 |
+
},
|
| 1815 |
+
{
|
| 1816 |
+
"epoch": 1.2692063492063492,
|
| 1817 |
+
"eval_logits/chosen": 5.721224784851074,
|
| 1818 |
+
"eval_logits/rejected": 5.573430061340332,
|
| 1819 |
+
"eval_logps/chosen": -3.8567862510681152,
|
| 1820 |
+
"eval_logps/rejected": -6.927263259887695,
|
| 1821 |
+
"eval_loss": 0.26795223355293274,
|
| 1822 |
+
"eval_rewards/accuracies": 0.9375,
|
| 1823 |
+
"eval_rewards/chosen": -5.785179615020752,
|
| 1824 |
+
"eval_rewards/margins": 4.605715751647949,
|
| 1825 |
+
"eval_rewards/rejected": -10.390894889831543,
|
| 1826 |
+
"eval_runtime": 18.3572,
|
| 1827 |
+
"eval_samples_per_second": 27.782,
|
| 1828 |
+
"eval_steps_per_second": 3.486,
|
| 1829 |
+
"step": 1000
|
| 1830 |
+
},
|
| 1831 |
+
{
|
| 1832 |
+
"epoch": 1.2819047619047619,
|
| 1833 |
+
"grad_norm": 4.340210914611816,
|
| 1834 |
+
"learning_rate": 1.204932476567175e-06,
|
| 1835 |
+
"logits/chosen": 5.455284595489502,
|
| 1836 |
+
"logits/rejected": 5.007905006408691,
|
| 1837 |
+
"logps/chosen": -4.618704795837402,
|
| 1838 |
+
"logps/rejected": -7.544007778167725,
|
| 1839 |
+
"loss": 0.3142,
|
| 1840 |
+
"rewards/accuracies": 0.875,
|
| 1841 |
+
"rewards/chosen": -6.928057670593262,
|
| 1842 |
+
"rewards/margins": 4.387955188751221,
|
| 1843 |
+
"rewards/rejected": -11.316012382507324,
|
| 1844 |
+
"step": 1010
|
| 1845 |
+
},
|
| 1846 |
+
{
|
| 1847 |
+
"epoch": 1.2946031746031745,
|
| 1848 |
+
"grad_norm": 3.371431350708008,
|
| 1849 |
+
"learning_rate": 1.160433012552508e-06,
|
| 1850 |
+
"logits/chosen": 5.39217472076416,
|
| 1851 |
+
"logits/rejected": 5.3231329917907715,
|
| 1852 |
+
"logps/chosen": -3.9590423107147217,
|
| 1853 |
+
"logps/rejected": -7.2238945960998535,
|
| 1854 |
+
"loss": 0.3226,
|
| 1855 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1856 |
+
"rewards/chosen": -5.938564300537109,
|
| 1857 |
+
"rewards/margins": 4.897278785705566,
|
| 1858 |
+
"rewards/rejected": -10.835843086242676,
|
| 1859 |
+
"step": 1020
|
| 1860 |
+
},
|
| 1861 |
+
{
|
| 1862 |
+
"epoch": 1.3073015873015872,
|
| 1863 |
+
"grad_norm": 3.215897560119629,
|
| 1864 |
+
"learning_rate": 1.11652112689164e-06,
|
| 1865 |
+
"logits/chosen": 5.496011734008789,
|
| 1866 |
+
"logits/rejected": 5.371179580688477,
|
| 1867 |
+
"logps/chosen": -4.006840705871582,
|
| 1868 |
+
"logps/rejected": -6.894573211669922,
|
| 1869 |
+
"loss": 0.2998,
|
| 1870 |
+
"rewards/accuracies": 0.875,
|
| 1871 |
+
"rewards/chosen": -6.0102620124816895,
|
| 1872 |
+
"rewards/margins": 4.331597805023193,
|
| 1873 |
+
"rewards/rejected": -10.341859817504883,
|
| 1874 |
+
"step": 1030
|
| 1875 |
+
},
|
| 1876 |
+
{
|
| 1877 |
+
"epoch": 1.32,
|
| 1878 |
+
"grad_norm": 2.8000078201293945,
|
| 1879 |
+
"learning_rate": 1.073216080788921e-06,
|
| 1880 |
+
"logits/chosen": 6.246156215667725,
|
| 1881 |
+
"logits/rejected": 5.520369052886963,
|
| 1882 |
+
"logps/chosen": -4.688235282897949,
|
| 1883 |
+
"logps/rejected": -8.393880844116211,
|
| 1884 |
+
"loss": 0.2691,
|
| 1885 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 1886 |
+
"rewards/chosen": -7.032352447509766,
|
| 1887 |
+
"rewards/margins": 5.558468341827393,
|
| 1888 |
+
"rewards/rejected": -12.590821266174316,
|
| 1889 |
+
"step": 1040
|
| 1890 |
+
},
|
| 1891 |
+
{
|
| 1892 |
+
"epoch": 1.3326984126984127,
|
| 1893 |
+
"grad_norm": 3.0592737197875977,
|
| 1894 |
+
"learning_rate": 1.0305368692688175e-06,
|
| 1895 |
+
"logits/chosen": 5.306058406829834,
|
| 1896 |
+
"logits/rejected": 5.3769707679748535,
|
| 1897 |
+
"logps/chosen": -4.188882350921631,
|
| 1898 |
+
"logps/rejected": -6.395564079284668,
|
| 1899 |
+
"loss": 0.3026,
|
| 1900 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 1901 |
+
"rewards/chosen": -6.283323764801025,
|
| 1902 |
+
"rewards/margins": 3.3100223541259766,
|
| 1903 |
+
"rewards/rejected": -9.593345642089844,
|
| 1904 |
+
"step": 1050
|
| 1905 |
+
},
|
| 1906 |
+
{
|
| 1907 |
+
"epoch": 1.3326984126984127,
|
| 1908 |
+
"eval_logits/chosen": 5.6976423263549805,
|
| 1909 |
+
"eval_logits/rejected": 5.525759696960449,
|
| 1910 |
+
"eval_logps/chosen": -3.8439218997955322,
|
| 1911 |
+
"eval_logps/rejected": -6.9627227783203125,
|
| 1912 |
+
"eval_loss": 0.26340675354003906,
|
| 1913 |
+
"eval_rewards/accuracies": 0.9375,
|
| 1914 |
+
"eval_rewards/chosen": -5.76588249206543,
|
| 1915 |
+
"eval_rewards/margins": 4.678201675415039,
|
| 1916 |
+
"eval_rewards/rejected": -10.444085121154785,
|
| 1917 |
+
"eval_runtime": 18.3435,
|
| 1918 |
+
"eval_samples_per_second": 27.803,
|
| 1919 |
+
"eval_steps_per_second": 3.489,
|
| 1920 |
+
"step": 1050
|
| 1921 |
+
},
|
| 1922 |
+
{
|
| 1923 |
+
"epoch": 1.3453968253968254,
|
| 1924 |
+
"grad_norm": 3.524646759033203,
|
| 1925 |
+
"learning_rate": 9.88502212844063e-07,
|
| 1926 |
+
"logits/chosen": 5.86399507522583,
|
| 1927 |
+
"logits/rejected": 5.444242000579834,
|
| 1928 |
+
"logps/chosen": -4.269190788269043,
|
| 1929 |
+
"logps/rejected": -7.391558647155762,
|
| 1930 |
+
"loss": 0.2872,
|
| 1931 |
+
"rewards/accuracies": 0.9375,
|
| 1932 |
+
"rewards/chosen": -6.403787136077881,
|
| 1933 |
+
"rewards/margins": 4.68355131149292,
|
| 1934 |
+
"rewards/rejected": -11.087339401245117,
|
| 1935 |
+
"step": 1060
|
| 1936 |
+
},
|
| 1937 |
+
{
|
| 1938 |
+
"epoch": 1.358095238095238,
|
| 1939 |
+
"grad_norm": 3.6155037879943848,
|
| 1940 |
+
"learning_rate": 9.471305493042243e-07,
|
| 1941 |
+
"logits/chosen": 6.007330894470215,
|
| 1942 |
+
"logits/rejected": 5.697009563446045,
|
| 1943 |
+
"logps/chosen": -4.269715785980225,
|
| 1944 |
+
"logps/rejected": -7.014040470123291,
|
| 1945 |
+
"loss": 0.2794,
|
| 1946 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1947 |
+
"rewards/chosen": -6.404573917388916,
|
| 1948 |
+
"rewards/margins": 4.116486072540283,
|
| 1949 |
+
"rewards/rejected": -10.5210599899292,
|
| 1950 |
+
"step": 1070
|
| 1951 |
+
},
|
| 1952 |
+
{
|
| 1953 |
+
"epoch": 1.370793650793651,
|
| 1954 |
+
"grad_norm": 4.943655014038086,
|
| 1955 |
+
"learning_rate": 9.064400256282757e-07,
|
| 1956 |
+
"logits/chosen": 5.4075846672058105,
|
| 1957 |
+
"logits/rejected": 5.163808345794678,
|
| 1958 |
+
"logps/chosen": -3.780120849609375,
|
| 1959 |
+
"logps/rejected": -6.776060581207275,
|
| 1960 |
+
"loss": 0.3154,
|
| 1961 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1962 |
+
"rewards/chosen": -5.6701812744140625,
|
| 1963 |
+
"rewards/margins": 4.4939093589782715,
|
| 1964 |
+
"rewards/rejected": -10.164091110229492,
|
| 1965 |
+
"step": 1080
|
| 1966 |
+
},
|
| 1967 |
+
{
|
| 1968 |
+
"epoch": 1.3834920634920636,
|
| 1969 |
+
"grad_norm": 3.8432955741882324,
|
| 1970 |
+
"learning_rate": 8.664484900247363e-07,
|
| 1971 |
+
"logits/chosen": 6.4478631019592285,
|
| 1972 |
+
"logits/rejected": 5.634939193725586,
|
| 1973 |
+
"logps/chosen": -4.478579521179199,
|
| 1974 |
+
"logps/rejected": -7.522899627685547,
|
| 1975 |
+
"loss": 0.3541,
|
| 1976 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 1977 |
+
"rewards/chosen": -6.717869758605957,
|
| 1978 |
+
"rewards/margins": 4.566478729248047,
|
| 1979 |
+
"rewards/rejected": -11.28434944152832,
|
| 1980 |
+
"step": 1090
|
| 1981 |
+
},
|
| 1982 |
+
{
|
| 1983 |
+
"epoch": 1.3961904761904762,
|
| 1984 |
+
"grad_norm": 5.117826461791992,
|
| 1985 |
+
"learning_rate": 8.271734841028553e-07,
|
| 1986 |
+
"logits/chosen": 5.143468379974365,
|
| 1987 |
+
"logits/rejected": 5.166830539703369,
|
| 1988 |
+
"logps/chosen": -3.964787721633911,
|
| 1989 |
+
"logps/rejected": -7.12085485458374,
|
| 1990 |
+
"loss": 0.3048,
|
| 1991 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1992 |
+
"rewards/chosen": -5.9471821784973145,
|
| 1993 |
+
"rewards/margins": 4.734100818634033,
|
| 1994 |
+
"rewards/rejected": -10.681283950805664,
|
| 1995 |
+
"step": 1100
|
| 1996 |
+
},
|
| 1997 |
+
{
|
| 1998 |
+
"epoch": 1.3961904761904762,
|
| 1999 |
+
"eval_logits/chosen": 5.741722583770752,
|
| 2000 |
+
"eval_logits/rejected": 5.563231468200684,
|
| 2001 |
+
"eval_logps/chosen": -3.90653395652771,
|
| 2002 |
+
"eval_logps/rejected": -7.086528778076172,
|
| 2003 |
+
"eval_loss": 0.2575596570968628,
|
| 2004 |
+
"eval_rewards/accuracies": 0.9375,
|
| 2005 |
+
"eval_rewards/chosen": -5.859801292419434,
|
| 2006 |
+
"eval_rewards/margins": 4.769991874694824,
|
| 2007 |
+
"eval_rewards/rejected": -10.629792213439941,
|
| 2008 |
+
"eval_runtime": 18.3447,
|
| 2009 |
+
"eval_samples_per_second": 27.801,
|
| 2010 |
+
"eval_steps_per_second": 3.489,
|
| 2011 |
+
"step": 1100
|
| 2012 |
+
},
|
| 2013 |
+
{
|
| 2014 |
+
"epoch": 1.4088888888888889,
|
| 2015 |
+
"grad_norm": 5.110182762145996,
|
| 2016 |
+
"learning_rate": 7.886322351782782e-07,
|
| 2017 |
+
"logits/chosen": 5.116390705108643,
|
| 2018 |
+
"logits/rejected": 4.859463691711426,
|
| 2019 |
+
"logps/chosen": -3.7181735038757324,
|
| 2020 |
+
"logps/rejected": -6.785368919372559,
|
| 2021 |
+
"loss": 0.2757,
|
| 2022 |
+
"rewards/accuracies": 0.949999988079071,
|
| 2023 |
+
"rewards/chosen": -5.5772600173950195,
|
| 2024 |
+
"rewards/margins": 4.600793838500977,
|
| 2025 |
+
"rewards/rejected": -10.178054809570312,
|
| 2026 |
+
"step": 1110
|
| 2027 |
+
},
|
| 2028 |
+
{
|
| 2029 |
+
"epoch": 1.4215873015873015,
|
| 2030 |
+
"grad_norm": 3.5725648403167725,
|
| 2031 |
+
"learning_rate": 7.508416487165862e-07,
|
| 2032 |
+
"logits/chosen": 5.8984761238098145,
|
| 2033 |
+
"logits/rejected": 5.381471633911133,
|
| 2034 |
+
"logps/chosen": -4.282950401306152,
|
| 2035 |
+
"logps/rejected": -7.246331691741943,
|
| 2036 |
+
"loss": 0.2697,
|
| 2037 |
+
"rewards/accuracies": 0.9375,
|
| 2038 |
+
"rewards/chosen": -6.424426078796387,
|
| 2039 |
+
"rewards/margins": 4.445071220397949,
|
| 2040 |
+
"rewards/rejected": -10.869497299194336,
|
| 2041 |
+
"step": 1120
|
| 2042 |
+
},
|
| 2043 |
+
{
|
| 2044 |
+
"epoch": 1.4342857142857142,
|
| 2045 |
+
"grad_norm": 4.783407688140869,
|
| 2046 |
+
"learning_rate": 7.138183009179922e-07,
|
| 2047 |
+
"logits/chosen": 5.907749176025391,
|
| 2048 |
+
"logits/rejected": 5.9079999923706055,
|
| 2049 |
+
"logps/chosen": -4.260853290557861,
|
| 2050 |
+
"logps/rejected": -7.2321672439575195,
|
| 2051 |
+
"loss": 0.2766,
|
| 2052 |
+
"rewards/accuracies": 0.949999988079071,
|
| 2053 |
+
"rewards/chosen": -6.391280174255371,
|
| 2054 |
+
"rewards/margins": 4.456971645355225,
|
| 2055 |
+
"rewards/rejected": -10.848252296447754,
|
| 2056 |
+
"step": 1130
|
| 2057 |
+
},
|
| 2058 |
+
{
|
| 2059 |
+
"epoch": 1.446984126984127,
|
| 2060 |
+
"grad_norm": 5.846556663513184,
|
| 2061 |
+
"learning_rate": 6.775784314464717e-07,
|
| 2062 |
+
"logits/chosen": 6.444309234619141,
|
| 2063 |
+
"logits/rejected": 5.971430778503418,
|
| 2064 |
+
"logps/chosen": -4.673151969909668,
|
| 2065 |
+
"logps/rejected": -7.5114336013793945,
|
| 2066 |
+
"loss": 0.2988,
|
| 2067 |
+
"rewards/accuracies": 0.887499988079071,
|
| 2068 |
+
"rewards/chosen": -7.00972843170166,
|
| 2069 |
+
"rewards/margins": 4.257421016693115,
|
| 2070 |
+
"rewards/rejected": -11.267148971557617,
|
| 2071 |
+
"step": 1140
|
| 2072 |
+
},
|
| 2073 |
+
{
|
| 2074 |
+
"epoch": 1.4596825396825397,
|
| 2075 |
+
"grad_norm": 2.654311418533325,
|
| 2076 |
+
"learning_rate": 6.421379363065142e-07,
|
| 2077 |
+
"logits/chosen": 6.01109504699707,
|
| 2078 |
+
"logits/rejected": 5.599249839782715,
|
| 2079 |
+
"logps/chosen": -4.461272239685059,
|
| 2080 |
+
"logps/rejected": -7.291404724121094,
|
| 2081 |
+
"loss": 0.2704,
|
| 2082 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 2083 |
+
"rewards/chosen": -6.691908836364746,
|
| 2084 |
+
"rewards/margins": 4.2451982498168945,
|
| 2085 |
+
"rewards/rejected": -10.937106132507324,
|
| 2086 |
+
"step": 1150
|
| 2087 |
+
},
|
| 2088 |
+
{
|
| 2089 |
+
"epoch": 1.4596825396825397,
|
| 2090 |
+
"eval_logits/chosen": 5.768884658813477,
|
| 2091 |
+
"eval_logits/rejected": 5.596267223358154,
|
| 2092 |
+
"eval_logps/chosen": -3.9878058433532715,
|
| 2093 |
+
"eval_logps/rejected": -7.195793151855469,
|
| 2094 |
+
"eval_loss": 0.25360986590385437,
|
| 2095 |
+
"eval_rewards/accuracies": 0.9375,
|
| 2096 |
+
"eval_rewards/chosen": -5.981709003448486,
|
| 2097 |
+
"eval_rewards/margins": 4.811980724334717,
|
| 2098 |
+
"eval_rewards/rejected": -10.793688774108887,
|
| 2099 |
+
"eval_runtime": 18.3358,
|
| 2100 |
+
"eval_samples_per_second": 27.814,
|
| 2101 |
+
"eval_steps_per_second": 3.49,
|
| 2102 |
+
"step": 1150
|
| 2103 |
+
},
|
| 2104 |
+
{
|
| 2105 |
+
"epoch": 1.4723809523809523,
|
| 2106 |
+
"grad_norm": 4.0658135414123535,
|
| 2107 |
+
"learning_rate": 6.075123608706093e-07,
|
| 2108 |
+
"logits/chosen": 5.25299072265625,
|
| 2109 |
+
"logits/rejected": 4.967761039733887,
|
| 2110 |
+
"logps/chosen": -4.026610851287842,
|
| 2111 |
+
"logps/rejected": -7.101304531097412,
|
| 2112 |
+
"loss": 0.2829,
|
| 2113 |
+
"rewards/accuracies": 0.9375,
|
| 2114 |
+
"rewards/chosen": -6.039916515350342,
|
| 2115 |
+
"rewards/margins": 4.612040042877197,
|
| 2116 |
+
"rewards/rejected": -10.651956558227539,
|
| 2117 |
+
"step": 1160
|
| 2118 |
+
},
|
| 2119 |
+
{
|
| 2120 |
+
"epoch": 1.485079365079365,
|
| 2121 |
+
"grad_norm": 3.9091060161590576,
|
| 2122 |
+
"learning_rate": 5.737168930605272e-07,
|
| 2123 |
+
"logits/chosen": 4.257780075073242,
|
| 2124 |
+
"logits/rejected": 4.387463569641113,
|
| 2125 |
+
"logps/chosen": -4.0905938148498535,
|
| 2126 |
+
"logps/rejected": -7.346035003662109,
|
| 2127 |
+
"loss": 0.2573,
|
| 2128 |
+
"rewards/accuracies": 0.9375,
|
| 2129 |
+
"rewards/chosen": -6.135890007019043,
|
| 2130 |
+
"rewards/margins": 4.8831610679626465,
|
| 2131 |
+
"rewards/rejected": -11.019052505493164,
|
| 2132 |
+
"step": 1170
|
| 2133 |
+
},
|
| 2134 |
+
{
|
| 2135 |
+
"epoch": 1.4977777777777779,
|
| 2136 |
+
"grad_norm": 4.522884368896484,
|
| 2137 |
+
"learning_rate": 5.407663566854008e-07,
|
| 2138 |
+
"logits/chosen": 6.337338447570801,
|
| 2139 |
+
"logits/rejected": 5.771265983581543,
|
| 2140 |
+
"logps/chosen": -3.8552029132843018,
|
| 2141 |
+
"logps/rejected": -6.726799011230469,
|
| 2142 |
+
"loss": 0.3046,
|
| 2143 |
+
"rewards/accuracies": 0.949999988079071,
|
| 2144 |
+
"rewards/chosen": -5.782804012298584,
|
| 2145 |
+
"rewards/margins": 4.307394981384277,
|
| 2146 |
+
"rewards/rejected": -10.090198516845703,
|
| 2147 |
+
"step": 1180
|
| 2148 |
+
},
|
| 2149 |
+
{
|
| 2150 |
+
"epoch": 1.5104761904761905,
|
| 2151 |
+
"grad_norm": 2.822293758392334,
|
| 2152 |
+
"learning_rate": 5.086752049395094e-07,
|
| 2153 |
+
"logits/chosen": 5.941995143890381,
|
| 2154 |
+
"logits/rejected": 5.562925815582275,
|
| 2155 |
+
"logps/chosen": -4.446758270263672,
|
| 2156 |
+
"logps/rejected": -7.707291603088379,
|
| 2157 |
+
"loss": 0.2863,
|
| 2158 |
+
"rewards/accuracies": 0.9375,
|
| 2159 |
+
"rewards/chosen": -6.67013692855835,
|
| 2160 |
+
"rewards/margins": 4.890799522399902,
|
| 2161 |
+
"rewards/rejected": -11.560935974121094,
|
| 2162 |
+
"step": 1190
|
| 2163 |
+
},
|
| 2164 |
+
{
|
| 2165 |
+
"epoch": 1.5231746031746032,
|
| 2166 |
+
"grad_norm": 3.358898401260376,
|
| 2167 |
+
"learning_rate": 4.774575140626317e-07,
|
| 2168 |
+
"logits/chosen": 6.258566856384277,
|
| 2169 |
+
"logits/rejected": 5.50495719909668,
|
| 2170 |
+
"logps/chosen": -5.074927806854248,
|
| 2171 |
+
"logps/rejected": -8.208864212036133,
|
| 2172 |
+
"loss": 0.274,
|
| 2173 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 2174 |
+
"rewards/chosen": -7.612391471862793,
|
| 2175 |
+
"rewards/margins": 4.70090389251709,
|
| 2176 |
+
"rewards/rejected": -12.313295364379883,
|
| 2177 |
+
"step": 1200
|
| 2178 |
+
},
|
| 2179 |
+
{
|
| 2180 |
+
"epoch": 1.5231746031746032,
|
| 2181 |
+
"eval_logits/chosen": 5.688702583312988,
|
| 2182 |
+
"eval_logits/rejected": 5.500263690948486,
|
| 2183 |
+
"eval_logps/chosen": -3.858718156814575,
|
| 2184 |
+
"eval_logps/rejected": -7.112142562866211,
|
| 2185 |
+
"eval_loss": 0.25126969814300537,
|
| 2186 |
+
"eval_rewards/accuracies": 0.9375,
|
| 2187 |
+
"eval_rewards/chosen": -5.788077354431152,
|
| 2188 |
+
"eval_rewards/margins": 4.880136489868164,
|
| 2189 |
+
"eval_rewards/rejected": -10.668213844299316,
|
| 2190 |
+
"eval_runtime": 18.3391,
|
| 2191 |
+
"eval_samples_per_second": 27.809,
|
| 2192 |
+
"eval_steps_per_second": 3.49,
|
| 2193 |
+
"step": 1200
|
| 2194 |
+
},
|
| 2195 |
+
{
|
| 2196 |
+
"epoch": 1.5358730158730158,
|
| 2197 |
+
"grad_norm": 2.7726964950561523,
|
| 2198 |
+
"learning_rate": 4.4712697716573994e-07,
|
| 2199 |
+
"logits/chosen": 6.500624656677246,
|
| 2200 |
+
"logits/rejected": 6.0488409996032715,
|
| 2201 |
+
"logps/chosen": -4.397651672363281,
|
| 2202 |
+
"logps/rejected": -7.040518760681152,
|
| 2203 |
+
"loss": 0.2878,
|
| 2204 |
+
"rewards/accuracies": 0.9375,
|
| 2205 |
+
"rewards/chosen": -6.596477508544922,
|
| 2206 |
+
"rewards/margins": 3.9643006324768066,
|
| 2207 |
+
"rewards/rejected": -10.560778617858887,
|
| 2208 |
+
"step": 1210
|
| 2209 |
+
},
|
| 2210 |
+
{
|
| 2211 |
+
"epoch": 1.5485714285714285,
|
| 2212 |
+
"grad_norm": 3.195594072341919,
|
| 2213 |
+
"learning_rate": 4.1769689822475147e-07,
|
| 2214 |
+
"logits/chosen": 5.61861515045166,
|
| 2215 |
+
"logits/rejected": 5.372084617614746,
|
| 2216 |
+
"logps/chosen": -3.832531452178955,
|
| 2217 |
+
"logps/rejected": -6.474020957946777,
|
| 2218 |
+
"loss": 0.2557,
|
| 2219 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 2220 |
+
"rewards/chosen": -5.748796463012695,
|
| 2221 |
+
"rewards/margins": 3.962235689163208,
|
| 2222 |
+
"rewards/rejected": -9.711031913757324,
|
| 2223 |
+
"step": 1220
|
| 2224 |
+
},
|
| 2225 |
+
{
|
| 2226 |
+
"epoch": 1.5612698412698411,
|
| 2227 |
+
"grad_norm": 3.5306906700134277,
|
| 2228 |
+
"learning_rate": 3.891801862449629e-07,
|
| 2229 |
+
"logits/chosen": 5.794755458831787,
|
| 2230 |
+
"logits/rejected": 5.1795549392700195,
|
| 2231 |
+
"logps/chosen": -3.9147841930389404,
|
| 2232 |
+
"logps/rejected": -7.290131568908691,
|
| 2233 |
+
"loss": 0.2745,
|
| 2234 |
+
"rewards/accuracies": 0.925000011920929,
|
| 2235 |
+
"rewards/chosen": -5.872176647186279,
|
| 2236 |
+
"rewards/margins": 5.063020706176758,
|
| 2237 |
+
"rewards/rejected": -10.935196876525879,
|
| 2238 |
+
"step": 1230
|
| 2239 |
+
},
|
| 2240 |
+
{
|
| 2241 |
+
"epoch": 1.573968253968254,
|
| 2242 |
+
"grad_norm": 4.873084545135498,
|
| 2243 |
+
"learning_rate": 3.615893495987335e-07,
|
| 2244 |
+
"logits/chosen": 5.19751501083374,
|
| 2245 |
+
"logits/rejected": 4.9404497146606445,
|
| 2246 |
+
"logps/chosen": -4.018169403076172,
|
| 2247 |
+
"logps/rejected": -6.874402046203613,
|
| 2248 |
+
"loss": 0.3246,
|
| 2249 |
+
"rewards/accuracies": 0.8500000238418579,
|
| 2250 |
+
"rewards/chosen": -6.0272536277771,
|
| 2251 |
+
"rewards/margins": 4.284348964691162,
|
| 2252 |
+
"rewards/rejected": -10.311602592468262,
|
| 2253 |
+
"step": 1240
|
| 2254 |
+
},
|
| 2255 |
+
{
|
| 2256 |
+
"epoch": 1.5866666666666667,
|
| 2257 |
+
"grad_norm": 4.694446086883545,
|
| 2258 |
+
"learning_rate": 3.3493649053890325e-07,
|
| 2259 |
+
"logits/chosen": 5.5149946212768555,
|
| 2260 |
+
"logits/rejected": 5.184729099273682,
|
| 2261 |
+
"logps/chosen": -4.031495571136475,
|
| 2262 |
+
"logps/rejected": -6.850281715393066,
|
| 2263 |
+
"loss": 0.2725,
|
| 2264 |
+
"rewards/accuracies": 0.925000011920929,
|
| 2265 |
+
"rewards/chosen": -6.047244071960449,
|
| 2266 |
+
"rewards/margins": 4.228178977966309,
|
| 2267 |
+
"rewards/rejected": -10.275423049926758,
|
| 2268 |
+
"step": 1250
|
| 2269 |
+
},
|
| 2270 |
+
{
|
| 2271 |
+
"epoch": 1.5866666666666667,
|
| 2272 |
+
"eval_logits/chosen": 5.6769232749938965,
|
| 2273 |
+
"eval_logits/rejected": 5.492671966552734,
|
| 2274 |
+
"eval_logps/chosen": -3.8999485969543457,
|
| 2275 |
+
"eval_logps/rejected": -7.1655073165893555,
|
| 2276 |
+
"eval_loss": 0.24976640939712524,
|
| 2277 |
+
"eval_rewards/accuracies": 0.9375,
|
| 2278 |
+
"eval_rewards/chosen": -5.8499226570129395,
|
| 2279 |
+
"eval_rewards/margins": 4.898338317871094,
|
| 2280 |
+
"eval_rewards/rejected": -10.748260498046875,
|
| 2281 |
+
"eval_runtime": 18.3399,
|
| 2282 |
+
"eval_samples_per_second": 27.808,
|
| 2283 |
+
"eval_steps_per_second": 3.49,
|
| 2284 |
+
"step": 1250
|
| 2285 |
+
}
|
| 2286 |
+
],
|
| 2287 |
+
"logging_steps": 10,
|
| 2288 |
+
"max_steps": 1500,
|
| 2289 |
+
"num_input_tokens_seen": 0,
|
| 2290 |
+
"num_train_epochs": 2,
|
| 2291 |
+
"save_steps": 50,
|
| 2292 |
+
"stateful_callbacks": {
|
| 2293 |
+
"TrainerControl": {
|
| 2294 |
+
"args": {
|
| 2295 |
+
"should_epoch_stop": false,
|
| 2296 |
+
"should_evaluate": false,
|
| 2297 |
+
"should_log": false,
|
| 2298 |
+
"should_save": true,
|
| 2299 |
+
"should_training_stop": false
|
| 2300 |
+
},
|
| 2301 |
+
"attributes": {}
|
| 2302 |
+
}
|
| 2303 |
+
},
|
| 2304 |
+
"total_flos": 3.015379311684223e+18,
|
| 2305 |
+
"train_batch_size": 1,
|
| 2306 |
+
"trial_name": null,
|
| 2307 |
+
"trial_params": null
|
| 2308 |
+
}
|
checkpoint-1250/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f6104f1a61db1df2c606ca77691d14a708e1aba41ef959437c3789e25b88c33b
|
| 3 |
+
size 7224
|
checkpoint-1250/zero_to_fp32.py
ADDED
|
@@ -0,0 +1,674 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
#!/usr/bin/env python
|
| 2 |
+
|
| 3 |
+
# Copyright (c) Microsoft Corporation.
|
| 4 |
+
# SPDX-License-Identifier: Apache-2.0
|
| 5 |
+
|
| 6 |
+
# DeepSpeed Team
|
| 7 |
+
|
| 8 |
+
# This script extracts fp32 consolidated weights from a zero 1, 2 and 3 DeepSpeed checkpoints. It gets
|
| 9 |
+
# copied into the top level checkpoint dir, so the user can easily do the conversion at any point in
|
| 10 |
+
# the future. Once extracted, the weights don't require DeepSpeed and can be used in any
|
| 11 |
+
# application.
|
| 12 |
+
#
|
| 13 |
+
# example:
|
| 14 |
+
# python zero_to_fp32.py . output_dir/
|
| 15 |
+
# or
|
| 16 |
+
# python zero_to_fp32.py . output_dir/ --safe_serialization
|
| 17 |
+
|
| 18 |
+
import argparse
|
| 19 |
+
import torch
|
| 20 |
+
import glob
|
| 21 |
+
import math
|
| 22 |
+
import os
|
| 23 |
+
import re
|
| 24 |
+
import json
|
| 25 |
+
from tqdm import tqdm
|
| 26 |
+
from collections import OrderedDict
|
| 27 |
+
from dataclasses import dataclass
|
| 28 |
+
|
| 29 |
+
# while this script doesn't use deepspeed to recover data, since the checkpoints are pickled with
|
| 30 |
+
# DeepSpeed data structures it has to be available in the current python environment.
|
| 31 |
+
from deepspeed.utils import logger
|
| 32 |
+
from deepspeed.checkpoint.constants import (DS_VERSION, OPTIMIZER_STATE_DICT, SINGLE_PARTITION_OF_FP32_GROUPS,
|
| 33 |
+
FP32_FLAT_GROUPS, ZERO_STAGE, PARTITION_COUNT, PARAM_SHAPES, BUFFER_NAMES,
|
| 34 |
+
FROZEN_PARAM_SHAPES, FROZEN_PARAM_FRAGMENTS)
|
| 35 |
+
|
| 36 |
+
|
| 37 |
+
@dataclass
|
| 38 |
+
class zero_model_state:
|
| 39 |
+
buffers: dict()
|
| 40 |
+
param_shapes: dict()
|
| 41 |
+
shared_params: list
|
| 42 |
+
ds_version: int
|
| 43 |
+
frozen_param_shapes: dict()
|
| 44 |
+
frozen_param_fragments: dict()
|
| 45 |
+
|
| 46 |
+
|
| 47 |
+
debug = 0
|
| 48 |
+
|
| 49 |
+
# load to cpu
|
| 50 |
+
device = torch.device('cpu')
|
| 51 |
+
|
| 52 |
+
|
| 53 |
+
def atoi(text):
|
| 54 |
+
return int(text) if text.isdigit() else text
|
| 55 |
+
|
| 56 |
+
|
| 57 |
+
def natural_keys(text):
|
| 58 |
+
'''
|
| 59 |
+
alist.sort(key=natural_keys) sorts in human order
|
| 60 |
+
http://nedbatchelder.com/blog/200712/human_sorting.html
|
| 61 |
+
(See Toothy's implementation in the comments)
|
| 62 |
+
'''
|
| 63 |
+
return [atoi(c) for c in re.split(r'(\d+)', text)]
|
| 64 |
+
|
| 65 |
+
|
| 66 |
+
def get_model_state_file(checkpoint_dir, zero_stage):
|
| 67 |
+
if not os.path.isdir(checkpoint_dir):
|
| 68 |
+
raise FileNotFoundError(f"Directory '{checkpoint_dir}' doesn't exist")
|
| 69 |
+
|
| 70 |
+
# there should be only one file
|
| 71 |
+
if zero_stage <= 2:
|
| 72 |
+
file = os.path.join(checkpoint_dir, "mp_rank_00_model_states.pt")
|
| 73 |
+
elif zero_stage == 3:
|
| 74 |
+
file = os.path.join(checkpoint_dir, "zero_pp_rank_0_mp_rank_00_model_states.pt")
|
| 75 |
+
|
| 76 |
+
if not os.path.exists(file):
|
| 77 |
+
raise FileNotFoundError(f"can't find model states file at '{file}'")
|
| 78 |
+
|
| 79 |
+
return file
|
| 80 |
+
|
| 81 |
+
|
| 82 |
+
def get_checkpoint_files(checkpoint_dir, glob_pattern):
|
| 83 |
+
# XXX: need to test that this simple glob rule works for multi-node setup too
|
| 84 |
+
ckpt_files = sorted(glob.glob(os.path.join(checkpoint_dir, glob_pattern)), key=natural_keys)
|
| 85 |
+
|
| 86 |
+
if len(ckpt_files) == 0:
|
| 87 |
+
raise FileNotFoundError(f"can't find {glob_pattern} files in directory '{checkpoint_dir}'")
|
| 88 |
+
|
| 89 |
+
return ckpt_files
|
| 90 |
+
|
| 91 |
+
|
| 92 |
+
def get_optim_files(checkpoint_dir):
|
| 93 |
+
return get_checkpoint_files(checkpoint_dir, "*_optim_states.pt")
|
| 94 |
+
|
| 95 |
+
|
| 96 |
+
def get_model_state_files(checkpoint_dir):
|
| 97 |
+
return get_checkpoint_files(checkpoint_dir, "*_model_states.pt")
|
| 98 |
+
|
| 99 |
+
|
| 100 |
+
def parse_model_states(files):
|
| 101 |
+
zero_model_states = []
|
| 102 |
+
for file in files:
|
| 103 |
+
state_dict = torch.load(file, map_location=device)
|
| 104 |
+
|
| 105 |
+
if BUFFER_NAMES not in state_dict:
|
| 106 |
+
raise ValueError(f"{file} is not a model state checkpoint")
|
| 107 |
+
buffer_names = state_dict[BUFFER_NAMES]
|
| 108 |
+
if debug:
|
| 109 |
+
print("Found buffers:", buffer_names)
|
| 110 |
+
|
| 111 |
+
# recover just the buffers while restoring them to fp32 if they were saved in fp16
|
| 112 |
+
buffers = {k: v.float() for k, v in state_dict["module"].items() if k in buffer_names}
|
| 113 |
+
param_shapes = state_dict[PARAM_SHAPES]
|
| 114 |
+
|
| 115 |
+
# collect parameters that are included in param_shapes
|
| 116 |
+
param_names = []
|
| 117 |
+
for s in param_shapes:
|
| 118 |
+
for name in s.keys():
|
| 119 |
+
param_names.append(name)
|
| 120 |
+
|
| 121 |
+
# update with frozen parameters
|
| 122 |
+
frozen_param_shapes = state_dict.get(FROZEN_PARAM_SHAPES, None)
|
| 123 |
+
if frozen_param_shapes is not None:
|
| 124 |
+
if debug:
|
| 125 |
+
print(f"Found frozen_param_shapes: {frozen_param_shapes}")
|
| 126 |
+
param_names += list(frozen_param_shapes.keys())
|
| 127 |
+
|
| 128 |
+
# handle shared params
|
| 129 |
+
shared_params = [[k, v] for k, v in state_dict["shared_params"].items()]
|
| 130 |
+
|
| 131 |
+
ds_version = state_dict.get(DS_VERSION, None)
|
| 132 |
+
|
| 133 |
+
frozen_param_fragments = state_dict.get(FROZEN_PARAM_FRAGMENTS, None)
|
| 134 |
+
|
| 135 |
+
z_model_state = zero_model_state(buffers=buffers,
|
| 136 |
+
param_shapes=param_shapes,
|
| 137 |
+
shared_params=shared_params,
|
| 138 |
+
ds_version=ds_version,
|
| 139 |
+
frozen_param_shapes=frozen_param_shapes,
|
| 140 |
+
frozen_param_fragments=frozen_param_fragments)
|
| 141 |
+
zero_model_states.append(z_model_state)
|
| 142 |
+
|
| 143 |
+
return zero_model_states
|
| 144 |
+
|
| 145 |
+
|
| 146 |
+
def parse_optim_states(files, ds_checkpoint_dir):
|
| 147 |
+
total_files = len(files)
|
| 148 |
+
state_dicts = []
|
| 149 |
+
for f in files:
|
| 150 |
+
state_dict = torch.load(f, map_location=device)
|
| 151 |
+
# immediately discard the potentially huge 2 optimizer states as we only care for fp32 master weights
|
| 152 |
+
# and also handle the case where it was already removed by another helper script
|
| 153 |
+
state_dict["optimizer_state_dict"].pop("optimizer_state_dict", None)
|
| 154 |
+
state_dicts.append(state_dict)
|
| 155 |
+
|
| 156 |
+
if not ZERO_STAGE in state_dicts[0][OPTIMIZER_STATE_DICT]:
|
| 157 |
+
raise ValueError(f"{files[0]} is not a zero checkpoint")
|
| 158 |
+
zero_stage = state_dicts[0][OPTIMIZER_STATE_DICT][ZERO_STAGE]
|
| 159 |
+
world_size = state_dicts[0][OPTIMIZER_STATE_DICT][PARTITION_COUNT]
|
| 160 |
+
|
| 161 |
+
# For ZeRO-2 each param group can have different partition_count as data parallelism for expert
|
| 162 |
+
# parameters can be different from data parallelism for non-expert parameters. So we can just
|
| 163 |
+
# use the max of the partition_count to get the dp world_size.
|
| 164 |
+
|
| 165 |
+
if type(world_size) is list:
|
| 166 |
+
world_size = max(world_size)
|
| 167 |
+
|
| 168 |
+
if world_size != total_files:
|
| 169 |
+
raise ValueError(
|
| 170 |
+
f"Expected {world_size} of '*_optim_states.pt' under '{ds_checkpoint_dir}' but found {total_files} files. "
|
| 171 |
+
"Possibly due to an overwrite of an old checkpoint, or a checkpoint didn't get saved by one or more processes."
|
| 172 |
+
)
|
| 173 |
+
|
| 174 |
+
# the groups are named differently in each stage
|
| 175 |
+
if zero_stage <= 2:
|
| 176 |
+
fp32_groups_key = SINGLE_PARTITION_OF_FP32_GROUPS
|
| 177 |
+
elif zero_stage == 3:
|
| 178 |
+
fp32_groups_key = FP32_FLAT_GROUPS
|
| 179 |
+
else:
|
| 180 |
+
raise ValueError(f"unknown zero stage {zero_stage}")
|
| 181 |
+
|
| 182 |
+
if zero_stage <= 2:
|
| 183 |
+
fp32_flat_groups = [state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key] for i in range(len(state_dicts))]
|
| 184 |
+
elif zero_stage == 3:
|
| 185 |
+
# if there is more than one param group, there will be multiple flattened tensors - one
|
| 186 |
+
# flattened tensor per group - for simplicity merge them into a single tensor
|
| 187 |
+
#
|
| 188 |
+
# XXX: could make the script more memory efficient for when there are multiple groups - it
|
| 189 |
+
# will require matching the sub-lists of param_shapes for each param group flattened tensor
|
| 190 |
+
|
| 191 |
+
fp32_flat_groups = [
|
| 192 |
+
torch.cat(state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key], 0) for i in range(len(state_dicts))
|
| 193 |
+
]
|
| 194 |
+
|
| 195 |
+
return zero_stage, world_size, fp32_flat_groups
|
| 196 |
+
|
| 197 |
+
|
| 198 |
+
def _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters):
|
| 199 |
+
"""
|
| 200 |
+
Returns fp32 state_dict reconstructed from ds checkpoint
|
| 201 |
+
|
| 202 |
+
Args:
|
| 203 |
+
- ``ds_checkpoint_dir``: path to the deepspeed checkpoint folder (where the optimizer files are)
|
| 204 |
+
|
| 205 |
+
"""
|
| 206 |
+
print(f"Processing zero checkpoint '{ds_checkpoint_dir}'")
|
| 207 |
+
|
| 208 |
+
optim_files = get_optim_files(ds_checkpoint_dir)
|
| 209 |
+
zero_stage, world_size, fp32_flat_groups = parse_optim_states(optim_files, ds_checkpoint_dir)
|
| 210 |
+
print(f"Detected checkpoint of type zero stage {zero_stage}, world_size: {world_size}")
|
| 211 |
+
|
| 212 |
+
model_files = get_model_state_files(ds_checkpoint_dir)
|
| 213 |
+
|
| 214 |
+
zero_model_states = parse_model_states(model_files)
|
| 215 |
+
print(f'Parsing checkpoint created by deepspeed=={zero_model_states[0].ds_version}')
|
| 216 |
+
|
| 217 |
+
if zero_stage <= 2:
|
| 218 |
+
return _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states,
|
| 219 |
+
exclude_frozen_parameters)
|
| 220 |
+
elif zero_stage == 3:
|
| 221 |
+
return _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states,
|
| 222 |
+
exclude_frozen_parameters)
|
| 223 |
+
|
| 224 |
+
|
| 225 |
+
def _zero2_merge_frozen_params(state_dict, zero_model_states):
|
| 226 |
+
if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
|
| 227 |
+
return
|
| 228 |
+
|
| 229 |
+
frozen_param_shapes = zero_model_states[0].frozen_param_shapes
|
| 230 |
+
frozen_param_fragments = zero_model_states[0].frozen_param_fragments
|
| 231 |
+
|
| 232 |
+
if debug:
|
| 233 |
+
num_elem = sum(s.numel() for s in frozen_param_shapes.values())
|
| 234 |
+
print(f'rank 0: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
|
| 235 |
+
|
| 236 |
+
wanted_params = len(frozen_param_shapes)
|
| 237 |
+
wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
|
| 238 |
+
avail_numel = sum([p.numel() for p in frozen_param_fragments.values()])
|
| 239 |
+
print(f'Frozen params: Have {avail_numel} numels to process.')
|
| 240 |
+
print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
|
| 241 |
+
|
| 242 |
+
total_params = 0
|
| 243 |
+
total_numel = 0
|
| 244 |
+
for name, shape in frozen_param_shapes.items():
|
| 245 |
+
total_params += 1
|
| 246 |
+
unpartitioned_numel = shape.numel()
|
| 247 |
+
total_numel += unpartitioned_numel
|
| 248 |
+
|
| 249 |
+
state_dict[name] = frozen_param_fragments[name]
|
| 250 |
+
|
| 251 |
+
if debug:
|
| 252 |
+
print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
|
| 253 |
+
|
| 254 |
+
print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
|
| 255 |
+
|
| 256 |
+
|
| 257 |
+
def _has_callable(obj, fn):
|
| 258 |
+
attr = getattr(obj, fn, None)
|
| 259 |
+
return callable(attr)
|
| 260 |
+
|
| 261 |
+
|
| 262 |
+
def _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
|
| 263 |
+
param_shapes = zero_model_states[0].param_shapes
|
| 264 |
+
|
| 265 |
+
# Reconstruction protocol:
|
| 266 |
+
#
|
| 267 |
+
# XXX: document this
|
| 268 |
+
|
| 269 |
+
if debug:
|
| 270 |
+
for i in range(world_size):
|
| 271 |
+
for j in range(len(fp32_flat_groups[0])):
|
| 272 |
+
print(f"{FP32_FLAT_GROUPS}[{i}][{j}].shape={fp32_flat_groups[i][j].shape}")
|
| 273 |
+
|
| 274 |
+
# XXX: memory usage doubles here (zero2)
|
| 275 |
+
num_param_groups = len(fp32_flat_groups[0])
|
| 276 |
+
merged_single_partition_of_fp32_groups = []
|
| 277 |
+
for i in range(num_param_groups):
|
| 278 |
+
merged_partitions = [sd[i] for sd in fp32_flat_groups]
|
| 279 |
+
full_single_fp32_vector = torch.cat(merged_partitions, 0)
|
| 280 |
+
merged_single_partition_of_fp32_groups.append(full_single_fp32_vector)
|
| 281 |
+
avail_numel = sum(
|
| 282 |
+
[full_single_fp32_vector.numel() for full_single_fp32_vector in merged_single_partition_of_fp32_groups])
|
| 283 |
+
|
| 284 |
+
if debug:
|
| 285 |
+
wanted_params = sum([len(shapes) for shapes in param_shapes])
|
| 286 |
+
wanted_numel = sum([sum(shape.numel() for shape in shapes.values()) for shapes in param_shapes])
|
| 287 |
+
# not asserting if there is a mismatch due to possible padding
|
| 288 |
+
print(f"Have {avail_numel} numels to process.")
|
| 289 |
+
print(f"Need {wanted_numel} numels in {wanted_params} params.")
|
| 290 |
+
|
| 291 |
+
# params
|
| 292 |
+
# XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
|
| 293 |
+
# out-of-core computing solution
|
| 294 |
+
total_numel = 0
|
| 295 |
+
total_params = 0
|
| 296 |
+
for shapes, full_single_fp32_vector in zip(param_shapes, merged_single_partition_of_fp32_groups):
|
| 297 |
+
offset = 0
|
| 298 |
+
avail_numel = full_single_fp32_vector.numel()
|
| 299 |
+
for name, shape in shapes.items():
|
| 300 |
+
|
| 301 |
+
unpartitioned_numel = shape.numel() if _has_callable(shape, 'numel') else math.prod(shape)
|
| 302 |
+
total_numel += unpartitioned_numel
|
| 303 |
+
total_params += 1
|
| 304 |
+
|
| 305 |
+
if debug:
|
| 306 |
+
print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
|
| 307 |
+
state_dict[name] = full_single_fp32_vector.narrow(0, offset, unpartitioned_numel).view(shape)
|
| 308 |
+
offset += unpartitioned_numel
|
| 309 |
+
|
| 310 |
+
# Z2 started to align to 2*world_size to improve nccl performance. Therefore both offset and
|
| 311 |
+
# avail_numel can differ by anywhere between 0..2*world_size. Due to two unrelated complex
|
| 312 |
+
# paddings performed in the code it's almost impossible to predict the exact numbers w/o the
|
| 313 |
+
# live optimizer object, so we are checking that the numbers are within the right range
|
| 314 |
+
align_to = 2 * world_size
|
| 315 |
+
|
| 316 |
+
def zero2_align(x):
|
| 317 |
+
return align_to * math.ceil(x / align_to)
|
| 318 |
+
|
| 319 |
+
if debug:
|
| 320 |
+
print(f"original offset={offset}, avail_numel={avail_numel}")
|
| 321 |
+
|
| 322 |
+
offset = zero2_align(offset)
|
| 323 |
+
avail_numel = zero2_align(avail_numel)
|
| 324 |
+
|
| 325 |
+
if debug:
|
| 326 |
+
print(f"aligned offset={offset}, avail_numel={avail_numel}")
|
| 327 |
+
|
| 328 |
+
# Sanity check
|
| 329 |
+
if offset != avail_numel:
|
| 330 |
+
raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
|
| 331 |
+
|
| 332 |
+
print(f"Reconstructed fp32 state dict with {total_params} params {total_numel} elements")
|
| 333 |
+
|
| 334 |
+
|
| 335 |
+
def _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states,
|
| 336 |
+
exclude_frozen_parameters):
|
| 337 |
+
state_dict = OrderedDict()
|
| 338 |
+
|
| 339 |
+
# buffers
|
| 340 |
+
buffers = zero_model_states[0].buffers
|
| 341 |
+
state_dict.update(buffers)
|
| 342 |
+
if debug:
|
| 343 |
+
print(f"added {len(buffers)} buffers")
|
| 344 |
+
|
| 345 |
+
if not exclude_frozen_parameters:
|
| 346 |
+
_zero2_merge_frozen_params(state_dict, zero_model_states)
|
| 347 |
+
|
| 348 |
+
_zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
|
| 349 |
+
|
| 350 |
+
# recover shared parameters
|
| 351 |
+
for pair in zero_model_states[0].shared_params:
|
| 352 |
+
if pair[1] in state_dict:
|
| 353 |
+
state_dict[pair[0]] = state_dict[pair[1]]
|
| 354 |
+
|
| 355 |
+
return state_dict
|
| 356 |
+
|
| 357 |
+
|
| 358 |
+
def zero3_partitioned_param_info(unpartitioned_numel, world_size):
|
| 359 |
+
remainder = unpartitioned_numel % world_size
|
| 360 |
+
padding_numel = (world_size - remainder) if remainder else 0
|
| 361 |
+
partitioned_numel = math.ceil(unpartitioned_numel / world_size)
|
| 362 |
+
return partitioned_numel, padding_numel
|
| 363 |
+
|
| 364 |
+
|
| 365 |
+
def _zero3_merge_frozen_params(state_dict, world_size, zero_model_states):
|
| 366 |
+
if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
|
| 367 |
+
return
|
| 368 |
+
|
| 369 |
+
if debug:
|
| 370 |
+
for i in range(world_size):
|
| 371 |
+
num_elem = sum(s.numel() for s in zero_model_states[i].frozen_param_fragments.values())
|
| 372 |
+
print(f'rank {i}: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
|
| 373 |
+
|
| 374 |
+
frozen_param_shapes = zero_model_states[0].frozen_param_shapes
|
| 375 |
+
wanted_params = len(frozen_param_shapes)
|
| 376 |
+
wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
|
| 377 |
+
avail_numel = sum([p.numel() for p in zero_model_states[0].frozen_param_fragments.values()]) * world_size
|
| 378 |
+
print(f'Frozen params: Have {avail_numel} numels to process.')
|
| 379 |
+
print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
|
| 380 |
+
|
| 381 |
+
total_params = 0
|
| 382 |
+
total_numel = 0
|
| 383 |
+
for name, shape in zero_model_states[0].frozen_param_shapes.items():
|
| 384 |
+
total_params += 1
|
| 385 |
+
unpartitioned_numel = shape.numel()
|
| 386 |
+
total_numel += unpartitioned_numel
|
| 387 |
+
|
| 388 |
+
param_frags = tuple(model_state.frozen_param_fragments[name] for model_state in zero_model_states)
|
| 389 |
+
state_dict[name] = torch.cat(param_frags, 0).narrow(0, 0, unpartitioned_numel).view(shape)
|
| 390 |
+
|
| 391 |
+
partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
|
| 392 |
+
|
| 393 |
+
if debug:
|
| 394 |
+
print(
|
| 395 |
+
f"Frozen params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
|
| 396 |
+
)
|
| 397 |
+
|
| 398 |
+
print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
|
| 399 |
+
|
| 400 |
+
|
| 401 |
+
def _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
|
| 402 |
+
param_shapes = zero_model_states[0].param_shapes
|
| 403 |
+
avail_numel = fp32_flat_groups[0].numel() * world_size
|
| 404 |
+
# Reconstruction protocol: For zero3 we need to zip the partitions together at boundary of each
|
| 405 |
+
# param, re-consolidating each param, while dealing with padding if any
|
| 406 |
+
|
| 407 |
+
# merge list of dicts, preserving order
|
| 408 |
+
param_shapes = {k: v for d in param_shapes for k, v in d.items()}
|
| 409 |
+
|
| 410 |
+
if debug:
|
| 411 |
+
for i in range(world_size):
|
| 412 |
+
print(f"{FP32_FLAT_GROUPS}[{i}].shape={fp32_flat_groups[i].shape}")
|
| 413 |
+
|
| 414 |
+
wanted_params = len(param_shapes)
|
| 415 |
+
wanted_numel = sum(shape.numel() for shape in param_shapes.values())
|
| 416 |
+
# not asserting if there is a mismatch due to possible padding
|
| 417 |
+
avail_numel = fp32_flat_groups[0].numel() * world_size
|
| 418 |
+
print(f"Trainable params: Have {avail_numel} numels to process.")
|
| 419 |
+
print(f"Trainable params: Need {wanted_numel} numels in {wanted_params} params.")
|
| 420 |
+
|
| 421 |
+
# params
|
| 422 |
+
# XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
|
| 423 |
+
# out-of-core computing solution
|
| 424 |
+
offset = 0
|
| 425 |
+
total_numel = 0
|
| 426 |
+
total_params = 0
|
| 427 |
+
for name, shape in tqdm(param_shapes.items(), desc='Gathering Sharded Weights'):
|
| 428 |
+
unpartitioned_numel = shape.numel()
|
| 429 |
+
total_numel += unpartitioned_numel
|
| 430 |
+
total_params += 1
|
| 431 |
+
partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
|
| 432 |
+
|
| 433 |
+
if debug:
|
| 434 |
+
print(
|
| 435 |
+
f"Trainable params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
|
| 436 |
+
)
|
| 437 |
+
|
| 438 |
+
# XXX: memory usage doubles here
|
| 439 |
+
state_dict[name] = torch.cat(
|
| 440 |
+
tuple(fp32_flat_groups[i].narrow(0, offset, partitioned_numel) for i in range(world_size)),
|
| 441 |
+
0).narrow(0, 0, unpartitioned_numel).view(shape)
|
| 442 |
+
offset += partitioned_numel
|
| 443 |
+
|
| 444 |
+
offset *= world_size
|
| 445 |
+
|
| 446 |
+
# Sanity check
|
| 447 |
+
if offset != avail_numel:
|
| 448 |
+
raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
|
| 449 |
+
|
| 450 |
+
print(f"Reconstructed Trainable fp32 state dict with {total_params} params {total_numel} elements")
|
| 451 |
+
|
| 452 |
+
|
| 453 |
+
def _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states,
|
| 454 |
+
exclude_frozen_parameters):
|
| 455 |
+
state_dict = OrderedDict()
|
| 456 |
+
|
| 457 |
+
# buffers
|
| 458 |
+
buffers = zero_model_states[0].buffers
|
| 459 |
+
state_dict.update(buffers)
|
| 460 |
+
if debug:
|
| 461 |
+
print(f"added {len(buffers)} buffers")
|
| 462 |
+
|
| 463 |
+
if not exclude_frozen_parameters:
|
| 464 |
+
_zero3_merge_frozen_params(state_dict, world_size, zero_model_states)
|
| 465 |
+
|
| 466 |
+
_zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
|
| 467 |
+
|
| 468 |
+
# recover shared parameters
|
| 469 |
+
for pair in zero_model_states[0].shared_params:
|
| 470 |
+
if pair[1] in state_dict:
|
| 471 |
+
state_dict[pair[0]] = state_dict[pair[1]]
|
| 472 |
+
|
| 473 |
+
return state_dict
|
| 474 |
+
|
| 475 |
+
|
| 476 |
+
def get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag=None, exclude_frozen_parameters=False):
|
| 477 |
+
"""
|
| 478 |
+
Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated state_dict that can be loaded with
|
| 479 |
+
``load_state_dict()`` and used for training without DeepSpeed or shared with others, for example
|
| 480 |
+
via a model hub.
|
| 481 |
+
|
| 482 |
+
Args:
|
| 483 |
+
- ``checkpoint_dir``: path to the desired checkpoint folder
|
| 484 |
+
- ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in 'latest' file. e.g., ``global_step14``
|
| 485 |
+
- ``exclude_frozen_parameters``: exclude frozen parameters
|
| 486 |
+
|
| 487 |
+
Returns:
|
| 488 |
+
- pytorch ``state_dict``
|
| 489 |
+
|
| 490 |
+
Note: this approach may not work if your application doesn't have sufficient free CPU memory and
|
| 491 |
+
you may need to use the offline approach using the ``zero_to_fp32.py`` script that is saved with
|
| 492 |
+
the checkpoint.
|
| 493 |
+
|
| 494 |
+
A typical usage might be ::
|
| 495 |
+
|
| 496 |
+
from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint
|
| 497 |
+
# do the training and checkpoint saving
|
| 498 |
+
state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir) # already on cpu
|
| 499 |
+
model = model.cpu() # move to cpu
|
| 500 |
+
model.load_state_dict(state_dict)
|
| 501 |
+
# submit to model hub or save the model to share with others
|
| 502 |
+
|
| 503 |
+
In this example the ``model`` will no longer be usable in the deepspeed context of the same
|
| 504 |
+
application. i.e. you will need to re-initialize the deepspeed engine, since
|
| 505 |
+
``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
|
| 506 |
+
|
| 507 |
+
If you want it all done for you, use ``load_state_dict_from_zero_checkpoint`` instead.
|
| 508 |
+
|
| 509 |
+
"""
|
| 510 |
+
if tag is None:
|
| 511 |
+
latest_path = os.path.join(checkpoint_dir, 'latest')
|
| 512 |
+
if os.path.isfile(latest_path):
|
| 513 |
+
with open(latest_path, 'r') as fd:
|
| 514 |
+
tag = fd.read().strip()
|
| 515 |
+
else:
|
| 516 |
+
raise ValueError(f"Unable to find 'latest' file at {latest_path}")
|
| 517 |
+
|
| 518 |
+
ds_checkpoint_dir = os.path.join(checkpoint_dir, tag)
|
| 519 |
+
|
| 520 |
+
if not os.path.isdir(ds_checkpoint_dir):
|
| 521 |
+
raise FileNotFoundError(f"Directory '{ds_checkpoint_dir}' doesn't exist")
|
| 522 |
+
|
| 523 |
+
return _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters)
|
| 524 |
+
|
| 525 |
+
|
| 526 |
+
def convert_zero_checkpoint_to_fp32_state_dict(checkpoint_dir,
|
| 527 |
+
output_dir,
|
| 528 |
+
max_shard_size="5GB",
|
| 529 |
+
safe_serialization=False,
|
| 530 |
+
tag=None,
|
| 531 |
+
exclude_frozen_parameters=False):
|
| 532 |
+
"""
|
| 533 |
+
Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` file that can be
|
| 534 |
+
loaded with ``torch.load(file)`` + ``load_state_dict()`` and used for training without DeepSpeed.
|
| 535 |
+
|
| 536 |
+
Args:
|
| 537 |
+
- ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
|
| 538 |
+
- ``output_dir``: directory to the pytorch fp32 state_dict output files
|
| 539 |
+
- ``max_shard_size``: the maximum size for a checkpoint before being sharded, default value is 5GB
|
| 540 |
+
- ``safe_serialization``: whether to save the model using `safetensors` or the traditional PyTorch way (that uses `pickle`).
|
| 541 |
+
- ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
|
| 542 |
+
- ``exclude_frozen_parameters``: exclude frozen parameters
|
| 543 |
+
"""
|
| 544 |
+
# Dependency pre-check
|
| 545 |
+
if safe_serialization:
|
| 546 |
+
try:
|
| 547 |
+
from safetensors.torch import save_file
|
| 548 |
+
except ImportError:
|
| 549 |
+
print('If you want to use `safe_serialization`, please `pip install safetensors`')
|
| 550 |
+
raise
|
| 551 |
+
if max_shard_size is not None:
|
| 552 |
+
try:
|
| 553 |
+
from huggingface_hub import split_torch_state_dict_into_shards
|
| 554 |
+
except ImportError:
|
| 555 |
+
print('If you want to use `max_shard_size`, please `pip install huggingface_hub`')
|
| 556 |
+
raise
|
| 557 |
+
|
| 558 |
+
# Convert zero checkpoint to state_dict
|
| 559 |
+
state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag, exclude_frozen_parameters)
|
| 560 |
+
|
| 561 |
+
# Shard the model if it is too big.
|
| 562 |
+
weights_name = "model.safetensors" if safe_serialization else "pytorch_model.bin"
|
| 563 |
+
if max_shard_size is not None:
|
| 564 |
+
filename_pattern = weights_name.replace(".bin", "{suffix}.bin").replace(".safetensors", "{suffix}.safetensors")
|
| 565 |
+
state_dict_split = split_torch_state_dict_into_shards(state_dict,
|
| 566 |
+
filename_pattern=filename_pattern,
|
| 567 |
+
max_shard_size=max_shard_size)
|
| 568 |
+
else:
|
| 569 |
+
from collections import namedtuple
|
| 570 |
+
StateDictSplit = namedtuple("StateDictSplit", ["is_sharded", "filename_to_tensors"])
|
| 571 |
+
state_dict_split = StateDictSplit(is_sharded=False,
|
| 572 |
+
filename_to_tensors={weights_name: list(state_dict.keys())})
|
| 573 |
+
|
| 574 |
+
# Save the model
|
| 575 |
+
filename_to_tensors = state_dict_split.filename_to_tensors.items()
|
| 576 |
+
for shard_file, tensors in tqdm(filename_to_tensors, desc="Saving checkpoint shards"):
|
| 577 |
+
shard = {tensor: state_dict[tensor].contiguous() for tensor in tensors}
|
| 578 |
+
output_path = os.path.join(output_dir, shard_file)
|
| 579 |
+
if safe_serialization:
|
| 580 |
+
save_file(shard, output_path, metadata={"format": "pt"})
|
| 581 |
+
else:
|
| 582 |
+
torch.save(shard, output_path)
|
| 583 |
+
|
| 584 |
+
# Save index if sharded
|
| 585 |
+
if state_dict_split.is_sharded:
|
| 586 |
+
index = {
|
| 587 |
+
"metadata": state_dict_split.metadata,
|
| 588 |
+
"weight_map": state_dict_split.tensor_to_filename,
|
| 589 |
+
}
|
| 590 |
+
save_index_file = "model.safetensors.index.json" if safe_serialization else "pytorch_model.bin.index.json"
|
| 591 |
+
save_index_file = os.path.join(output_dir, save_index_file)
|
| 592 |
+
with open(save_index_file, "w", encoding="utf-8") as f:
|
| 593 |
+
content = json.dumps(index, indent=2, sort_keys=True) + "\n"
|
| 594 |
+
f.write(content)
|
| 595 |
+
|
| 596 |
+
|
| 597 |
+
def load_state_dict_from_zero_checkpoint(model, checkpoint_dir, tag=None):
|
| 598 |
+
"""
|
| 599 |
+
1. Put the provided model to cpu
|
| 600 |
+
2. Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict``
|
| 601 |
+
3. Load it into the provided model
|
| 602 |
+
|
| 603 |
+
Args:
|
| 604 |
+
- ``model``: the model object to update
|
| 605 |
+
- ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
|
| 606 |
+
- ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
|
| 607 |
+
|
| 608 |
+
Returns:
|
| 609 |
+
- ``model`: modified model
|
| 610 |
+
|
| 611 |
+
Make sure you have plenty of CPU memory available before you call this function. If you don't
|
| 612 |
+
have enough use the ``zero_to_fp32.py`` utility to do the conversion. You will find it
|
| 613 |
+
conveniently placed for you in the checkpoint folder.
|
| 614 |
+
|
| 615 |
+
A typical usage might be ::
|
| 616 |
+
|
| 617 |
+
from deepspeed.utils.zero_to_fp32 import load_state_dict_from_zero_checkpoint
|
| 618 |
+
model = load_state_dict_from_zero_checkpoint(trainer.model, checkpoint_dir)
|
| 619 |
+
# submit to model hub or save the model to share with others
|
| 620 |
+
|
| 621 |
+
Note, that once this was run, the ``model`` will no longer be usable in the deepspeed context
|
| 622 |
+
of the same application. i.e. you will need to re-initialize the deepspeed engine, since
|
| 623 |
+
``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
|
| 624 |
+
|
| 625 |
+
"""
|
| 626 |
+
logger.info(f"Extracting fp32 weights")
|
| 627 |
+
state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag)
|
| 628 |
+
|
| 629 |
+
logger.info(f"Overwriting model with fp32 weights")
|
| 630 |
+
model = model.cpu()
|
| 631 |
+
model.load_state_dict(state_dict, strict=False)
|
| 632 |
+
|
| 633 |
+
return model
|
| 634 |
+
|
| 635 |
+
|
| 636 |
+
if __name__ == "__main__":
|
| 637 |
+
parser = argparse.ArgumentParser()
|
| 638 |
+
parser.add_argument("checkpoint_dir",
|
| 639 |
+
type=str,
|
| 640 |
+
help="path to the desired checkpoint folder, e.g., path/checkpoint-12")
|
| 641 |
+
parser.add_argument("output_dir",
|
| 642 |
+
type=str,
|
| 643 |
+
help="directory to the pytorch fp32 state_dict output files"
|
| 644 |
+
"(e.g. path/checkpoint-12-output/)")
|
| 645 |
+
parser.add_argument(
|
| 646 |
+
"--max_shard_size",
|
| 647 |
+
type=str,
|
| 648 |
+
default="5GB",
|
| 649 |
+
help="The maximum size for a checkpoint before being sharded. Checkpoints shard will then be each of size"
|
| 650 |
+
"lower than this size. If expressed as a string, needs to be digits followed by a unit (like `5MB`"
|
| 651 |
+
"We default it to 5GB in order for models to be able to run easily on free-tier google colab instances"
|
| 652 |
+
"without CPU OOM issues.")
|
| 653 |
+
parser.add_argument(
|
| 654 |
+
"--safe_serialization",
|
| 655 |
+
default=False,
|
| 656 |
+
action='store_true',
|
| 657 |
+
help="Whether to save the model using `safetensors` or the traditional PyTorch way (that uses `pickle`).")
|
| 658 |
+
parser.add_argument("-t",
|
| 659 |
+
"--tag",
|
| 660 |
+
type=str,
|
| 661 |
+
default=None,
|
| 662 |
+
help="checkpoint tag used as a unique identifier for checkpoint. e.g., global_step1")
|
| 663 |
+
parser.add_argument("--exclude_frozen_parameters", action='store_true', help="exclude frozen parameters")
|
| 664 |
+
parser.add_argument("-d", "--debug", action='store_true', help="enable debug")
|
| 665 |
+
args = parser.parse_args()
|
| 666 |
+
|
| 667 |
+
debug = args.debug
|
| 668 |
+
|
| 669 |
+
convert_zero_checkpoint_to_fp32_state_dict(args.checkpoint_dir,
|
| 670 |
+
args.output_dir,
|
| 671 |
+
max_shard_size=args.max_shard_size,
|
| 672 |
+
safe_serialization=args.safe_serialization,
|
| 673 |
+
tag=args.tag,
|
| 674 |
+
exclude_frozen_parameters=args.exclude_frozen_parameters)
|