Training in progress, step 1200, checkpoint
Browse files- checkpoint-1200/README.md +202 -0
- checkpoint-1200/adapter_config.json +34 -0
- checkpoint-1200/adapter_model.safetensors +3 -0
- checkpoint-1200/global_step1200/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt +3 -0
- checkpoint-1200/global_step1200/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt +3 -0
- checkpoint-1200/global_step1200/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt +3 -0
- checkpoint-1200/global_step1200/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt +3 -0
- checkpoint-1200/global_step1200/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt +3 -0
- checkpoint-1200/global_step1200/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt +3 -0
- checkpoint-1200/global_step1200/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt +3 -0
- checkpoint-1200/global_step1200/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt +3 -0
- checkpoint-1200/global_step1200/mp_rank_00_model_states.pt +3 -0
- checkpoint-1200/latest +1 -0
- checkpoint-1200/rng_state_0.pth +3 -0
- checkpoint-1200/rng_state_1.pth +3 -0
- checkpoint-1200/rng_state_2.pth +3 -0
- checkpoint-1200/rng_state_3.pth +3 -0
- checkpoint-1200/rng_state_4.pth +3 -0
- checkpoint-1200/rng_state_5.pth +3 -0
- checkpoint-1200/rng_state_6.pth +3 -0
- checkpoint-1200/rng_state_7.pth +3 -0
- checkpoint-1200/scheduler.pt +3 -0
- checkpoint-1200/special_tokens_map.json +30 -0
- checkpoint-1200/tokenizer.json +0 -0
- checkpoint-1200/tokenizer_config.json +133 -0
- checkpoint-1200/trainer_state.json +2217 -0
- checkpoint-1200/training_args.bin +3 -0
- checkpoint-1200/zero_to_fp32.py +674 -0
checkpoint-1200/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-1200/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 |
+
"down_proj",
|
| 29 |
+
"qkv_proj"
|
| 30 |
+
],
|
| 31 |
+
"task_type": "CAUSAL_LM",
|
| 32 |
+
"use_dora": false,
|
| 33 |
+
"use_rslora": false
|
| 34 |
+
}
|
checkpoint-1200/adapter_model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d938374b178488517a82c746688512fcb110abae712a4823316beac5d5b77af1
|
| 3 |
+
size 25200088
|
checkpoint-1200/global_step1200/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 sha256:93bbc426b26bd869db2c8c3e606764ef487a8581cb85356db63bb691ce089d42
|
| 3 |
+
size 18881328
|
checkpoint-1200/global_step1200/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:b4967c785ffdeaef21f357b78589119b0d2f0a93a8f518afb713cd15578a3464
|
| 3 |
+
size 18881328
|
checkpoint-1200/global_step1200/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:c83fbaf192e2e93079e719f6d0d10bc11825cb140b7d03ccf6ee37d1c9048701
|
| 3 |
+
size 18881328
|
checkpoint-1200/global_step1200/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:d94dbd77d96ec9cf09c30b8eb4cfc726427aeca4d871440cd7c74ec1bb8cdb4a
|
| 3 |
+
size 18881392
|
checkpoint-1200/global_step1200/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:d597054201aa09b4bdfc79a8959d24efda101575451b8c153bd039d9e45271e9
|
| 3 |
+
size 18881392
|
checkpoint-1200/global_step1200/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 sha256:6b2dc35a179967893c0d063beefcc2cd37ffd379796c94c1002791da3c88cebe
|
| 3 |
+
size 18881392
|
checkpoint-1200/global_step1200/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 sha256:1355cd56606d9518501f32a182a84494fba1249e0adaac7b1f342f9a361d204e
|
| 3 |
+
size 18881392
|
checkpoint-1200/global_step1200/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:aff360ce6c5d6cccdaf4edf06c4d280e1e8db160844e2bf0d060c68096d0b547
|
| 3 |
+
size 18881392
|
checkpoint-1200/global_step1200/mp_rank_00_model_states.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a80dbead0b4c3ad1821dbc111afe8fae86bb2e9f26c16f807f2ee5cafaf7a8b4
|
| 3 |
+
size 25379244
|
checkpoint-1200/latest
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
global_step1200
|
checkpoint-1200/rng_state_0.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5504b8d722b425f58bab6aedf9a43fc8129b02036307d31c7a21e224d2412ace
|
| 3 |
+
size 15984
|
checkpoint-1200/rng_state_1.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e555860fd7a2cfb8945f188f7232baf938ce622886881cc422b3eb0e7444eda4
|
| 3 |
+
size 15984
|
checkpoint-1200/rng_state_2.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ea0c272cf77c9504efaa077bfa8f9229d461c16d6641be0e57a7f20f9b761399
|
| 3 |
+
size 15984
|
checkpoint-1200/rng_state_3.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:921e0812b510be6ea788fc2c6aa7541f3ff4eb1bb3dd7c230340a35d8e1e764b
|
| 3 |
+
size 15984
|
checkpoint-1200/rng_state_4.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c3603a5403f0489f4be4d27720a0fa7e0fe0d08dbde5d58c1060cef37b9084d2
|
| 3 |
+
size 15984
|
checkpoint-1200/rng_state_5.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2b9f685b83b5545cc2db9c29e88184590e89acb7836b4bb92a6a1df01b4bf43f
|
| 3 |
+
size 15984
|
checkpoint-1200/rng_state_6.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5843202dd0ab5bc42fd0b6fa35e7cc2dca365d38fb379a2faf93bf274ef023e6
|
| 3 |
+
size 15984
|
checkpoint-1200/rng_state_7.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:79f72ffc2afb7672fc32ddd050c69181c2c0c16f8eac79a352eecb064fb5a9c7
|
| 3 |
+
size 15984
|
checkpoint-1200/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:35b4cc7927ca0f1b9b45cd00f72746408c82d953cb952c75e7569243d9fa3f0c
|
| 3 |
+
size 1064
|
checkpoint-1200/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-1200/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
checkpoint-1200/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-1200/trainer_state.json
ADDED
|
@@ -0,0 +1,2217 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.9687184661957619,
|
| 5 |
+
"eval_steps": 50,
|
| 6 |
+
"global_step": 1200,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.008072653884964682,
|
| 13 |
+
"grad_norm": 0.04926518723368645,
|
| 14 |
+
"learning_rate": 4.999451708687114e-06,
|
| 15 |
+
"logits/chosen": 14.542106628417969,
|
| 16 |
+
"logits/rejected": 14.864250183105469,
|
| 17 |
+
"logps/chosen": -0.2809702754020691,
|
| 18 |
+
"logps/rejected": -0.3013763725757599,
|
| 19 |
+
"loss": 0.9221,
|
| 20 |
+
"rewards/accuracies": 0.42500001192092896,
|
| 21 |
+
"rewards/chosen": -0.421455442905426,
|
| 22 |
+
"rewards/margins": 0.03060910664498806,
|
| 23 |
+
"rewards/rejected": -0.452064573764801,
|
| 24 |
+
"step": 10
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.016145307769929364,
|
| 28 |
+
"grad_norm": 0.056249432265758514,
|
| 29 |
+
"learning_rate": 4.997807075247147e-06,
|
| 30 |
+
"logits/chosen": 14.614748001098633,
|
| 31 |
+
"logits/rejected": 15.259109497070312,
|
| 32 |
+
"logps/chosen": -0.2828002870082855,
|
| 33 |
+
"logps/rejected": -0.3477819561958313,
|
| 34 |
+
"loss": 0.916,
|
| 35 |
+
"rewards/accuracies": 0.550000011920929,
|
| 36 |
+
"rewards/chosen": -0.4242004454135895,
|
| 37 |
+
"rewards/margins": 0.09747247397899628,
|
| 38 |
+
"rewards/rejected": -0.5216729044914246,
|
| 39 |
+
"step": 20
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.024217961654894045,
|
| 43 |
+
"grad_norm": 0.0664869099855423,
|
| 44 |
+
"learning_rate": 4.9950668210706795e-06,
|
| 45 |
+
"logits/chosen": 14.391204833984375,
|
| 46 |
+
"logits/rejected": 14.82734203338623,
|
| 47 |
+
"logps/chosen": -0.287629634141922,
|
| 48 |
+
"logps/rejected": -0.3329126834869385,
|
| 49 |
+
"loss": 0.914,
|
| 50 |
+
"rewards/accuracies": 0.48750001192092896,
|
| 51 |
+
"rewards/chosen": -0.4314444661140442,
|
| 52 |
+
"rewards/margins": 0.06792456656694412,
|
| 53 |
+
"rewards/rejected": -0.4993689954280853,
|
| 54 |
+
"step": 30
|
| 55 |
+
},
|
| 56 |
+
{
|
| 57 |
+
"epoch": 0.03229061553985873,
|
| 58 |
+
"grad_norm": 0.055584829300642014,
|
| 59 |
+
"learning_rate": 4.9912321481237616e-06,
|
| 60 |
+
"logits/chosen": 13.93278980255127,
|
| 61 |
+
"logits/rejected": 14.886846542358398,
|
| 62 |
+
"logps/chosen": -0.28155946731567383,
|
| 63 |
+
"logps/rejected": -0.3678051829338074,
|
| 64 |
+
"loss": 0.9273,
|
| 65 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 66 |
+
"rewards/chosen": -0.42233920097351074,
|
| 67 |
+
"rewards/margins": 0.12936851382255554,
|
| 68 |
+
"rewards/rejected": -0.5517078042030334,
|
| 69 |
+
"step": 40
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 0.04036326942482341,
|
| 73 |
+
"grad_norm": 0.07997103035449982,
|
| 74 |
+
"learning_rate": 4.986304738420684e-06,
|
| 75 |
+
"logits/chosen": 14.172693252563477,
|
| 76 |
+
"logits/rejected": 14.742494583129883,
|
| 77 |
+
"logps/chosen": -0.28225988149642944,
|
| 78 |
+
"logps/rejected": -0.34329456090927124,
|
| 79 |
+
"loss": 0.9081,
|
| 80 |
+
"rewards/accuracies": 0.48750001192092896,
|
| 81 |
+
"rewards/chosen": -0.42338982224464417,
|
| 82 |
+
"rewards/margins": 0.0915520042181015,
|
| 83 |
+
"rewards/rejected": -0.5149418115615845,
|
| 84 |
+
"step": 50
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 0.04036326942482341,
|
| 88 |
+
"eval_logits/chosen": 14.200166702270508,
|
| 89 |
+
"eval_logits/rejected": 14.817726135253906,
|
| 90 |
+
"eval_logps/chosen": -0.2625390887260437,
|
| 91 |
+
"eval_logps/rejected": -0.3458769917488098,
|
| 92 |
+
"eval_loss": 0.9080610275268555,
|
| 93 |
+
"eval_rewards/accuracies": 0.5544554591178894,
|
| 94 |
+
"eval_rewards/chosen": -0.39380866289138794,
|
| 95 |
+
"eval_rewards/margins": 0.1250067949295044,
|
| 96 |
+
"eval_rewards/rejected": -0.5188154578208923,
|
| 97 |
+
"eval_runtime": 29.8098,
|
| 98 |
+
"eval_samples_per_second": 26.87,
|
| 99 |
+
"eval_steps_per_second": 3.388,
|
| 100 |
+
"step": 50
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.04843592330978809,
|
| 104 |
+
"grad_norm": 0.06322095543146133,
|
| 105 |
+
"learning_rate": 4.980286753286196e-06,
|
| 106 |
+
"logits/chosen": 14.008010864257812,
|
| 107 |
+
"logits/rejected": 14.975939750671387,
|
| 108 |
+
"logps/chosen": -0.2697351574897766,
|
| 109 |
+
"logps/rejected": -0.3445274233818054,
|
| 110 |
+
"loss": 0.9195,
|
| 111 |
+
"rewards/accuracies": 0.5,
|
| 112 |
+
"rewards/chosen": -0.40460270643234253,
|
| 113 |
+
"rewards/margins": 0.1121884360909462,
|
| 114 |
+
"rewards/rejected": -0.5167912244796753,
|
| 115 |
+
"step": 60
|
| 116 |
+
},
|
| 117 |
+
{
|
| 118 |
+
"epoch": 0.056508577194752774,
|
| 119 |
+
"grad_norm": 0.12975476682186127,
|
| 120 |
+
"learning_rate": 4.973180832407471e-06,
|
| 121 |
+
"logits/chosen": 14.213134765625,
|
| 122 |
+
"logits/rejected": 14.889978408813477,
|
| 123 |
+
"logps/chosen": -0.2839818596839905,
|
| 124 |
+
"logps/rejected": -0.3382417559623718,
|
| 125 |
+
"loss": 0.9241,
|
| 126 |
+
"rewards/accuracies": 0.5375000238418579,
|
| 127 |
+
"rewards/chosen": -0.4259727895259857,
|
| 128 |
+
"rewards/margins": 0.08138985931873322,
|
| 129 |
+
"rewards/rejected": -0.5073626637458801,
|
| 130 |
+
"step": 70
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 0.06458123107971746,
|
| 134 |
+
"grad_norm": 0.08352109789848328,
|
| 135 |
+
"learning_rate": 4.964990092676263e-06,
|
| 136 |
+
"logits/chosen": 14.395146369934082,
|
| 137 |
+
"logits/rejected": 14.900177001953125,
|
| 138 |
+
"logps/chosen": -0.27067264914512634,
|
| 139 |
+
"logps/rejected": -0.3516673743724823,
|
| 140 |
+
"loss": 0.9251,
|
| 141 |
+
"rewards/accuracies": 0.637499988079071,
|
| 142 |
+
"rewards/chosen": -0.4060089588165283,
|
| 143 |
+
"rewards/margins": 0.12149210274219513,
|
| 144 |
+
"rewards/rejected": -0.5275009870529175,
|
| 145 |
+
"step": 80
|
| 146 |
+
},
|
| 147 |
+
{
|
| 148 |
+
"epoch": 0.07265388496468214,
|
| 149 |
+
"grad_norm": 0.0793827548623085,
|
| 150 |
+
"learning_rate": 4.9557181268217225e-06,
|
| 151 |
+
"logits/chosen": 13.905145645141602,
|
| 152 |
+
"logits/rejected": 14.897878646850586,
|
| 153 |
+
"logps/chosen": -0.26814645528793335,
|
| 154 |
+
"logps/rejected": -0.38061466813087463,
|
| 155 |
+
"loss": 0.9055,
|
| 156 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 157 |
+
"rewards/chosen": -0.4022197127342224,
|
| 158 |
+
"rewards/margins": 0.16870227456092834,
|
| 159 |
+
"rewards/rejected": -0.5709219574928284,
|
| 160 |
+
"step": 90
|
| 161 |
+
},
|
| 162 |
+
{
|
| 163 |
+
"epoch": 0.08072653884964683,
|
| 164 |
+
"grad_norm": 0.09212008118629456,
|
| 165 |
+
"learning_rate": 4.9453690018345144e-06,
|
| 166 |
+
"logits/chosen": 14.11627197265625,
|
| 167 |
+
"logits/rejected": 14.571977615356445,
|
| 168 |
+
"logps/chosen": -0.3130624294281006,
|
| 169 |
+
"logps/rejected": -0.35073375701904297,
|
| 170 |
+
"loss": 0.911,
|
| 171 |
+
"rewards/accuracies": 0.48750001192092896,
|
| 172 |
+
"rewards/chosen": -0.46959367394447327,
|
| 173 |
+
"rewards/margins": 0.05650699883699417,
|
| 174 |
+
"rewards/rejected": -0.5261006951332092,
|
| 175 |
+
"step": 100
|
| 176 |
+
},
|
| 177 |
+
{
|
| 178 |
+
"epoch": 0.08072653884964683,
|
| 179 |
+
"eval_logits/chosen": 13.845876693725586,
|
| 180 |
+
"eval_logits/rejected": 14.490789413452148,
|
| 181 |
+
"eval_logps/chosen": -0.2531408667564392,
|
| 182 |
+
"eval_logps/rejected": -0.3464036285877228,
|
| 183 |
+
"eval_loss": 0.8986235857009888,
|
| 184 |
+
"eval_rewards/accuracies": 0.5544554591178894,
|
| 185 |
+
"eval_rewards/chosen": -0.3797112703323364,
|
| 186 |
+
"eval_rewards/margins": 0.13989417254924774,
|
| 187 |
+
"eval_rewards/rejected": -0.5196054577827454,
|
| 188 |
+
"eval_runtime": 29.0914,
|
| 189 |
+
"eval_samples_per_second": 27.534,
|
| 190 |
+
"eval_steps_per_second": 3.472,
|
| 191 |
+
"step": 100
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.08879919273461151,
|
| 195 |
+
"grad_norm": 0.08554862439632416,
|
| 196 |
+
"learning_rate": 4.933947257182901e-06,
|
| 197 |
+
"logits/chosen": 13.416229248046875,
|
| 198 |
+
"logits/rejected": 14.582674026489258,
|
| 199 |
+
"logps/chosen": -0.2523443102836609,
|
| 200 |
+
"logps/rejected": -0.38751405477523804,
|
| 201 |
+
"loss": 0.8993,
|
| 202 |
+
"rewards/accuracies": 0.6875,
|
| 203 |
+
"rewards/chosen": -0.3785164952278137,
|
| 204 |
+
"rewards/margins": 0.20275457203388214,
|
| 205 |
+
"rewards/rejected": -0.5812710523605347,
|
| 206 |
+
"step": 110
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.09687184661957618,
|
| 210 |
+
"grad_norm": 0.29209578037261963,
|
| 211 |
+
"learning_rate": 4.921457902821578e-06,
|
| 212 |
+
"logits/chosen": 13.874654769897461,
|
| 213 |
+
"logits/rejected": 14.423624992370605,
|
| 214 |
+
"logps/chosen": -0.27300480008125305,
|
| 215 |
+
"logps/rejected": -0.3526575267314911,
|
| 216 |
+
"loss": 0.8942,
|
| 217 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 218 |
+
"rewards/chosen": -0.40950721502304077,
|
| 219 |
+
"rewards/margins": 0.11947910487651825,
|
| 220 |
+
"rewards/rejected": -0.5289863348007202,
|
| 221 |
+
"step": 120
|
| 222 |
+
},
|
| 223 |
+
{
|
| 224 |
+
"epoch": 0.10494450050454086,
|
| 225 |
+
"grad_norm": 0.1028478816151619,
|
| 226 |
+
"learning_rate": 4.907906416994146e-06,
|
| 227 |
+
"logits/chosen": 14.030723571777344,
|
| 228 |
+
"logits/rejected": 14.711235046386719,
|
| 229 |
+
"logps/chosen": -0.27410784363746643,
|
| 230 |
+
"logps/rejected": -0.3665519058704376,
|
| 231 |
+
"loss": 0.8922,
|
| 232 |
+
"rewards/accuracies": 0.550000011920929,
|
| 233 |
+
"rewards/chosen": -0.41116175055503845,
|
| 234 |
+
"rewards/margins": 0.1386660784482956,
|
| 235 |
+
"rewards/rejected": -0.5498278737068176,
|
| 236 |
+
"step": 130
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 0.11301715438950555,
|
| 240 |
+
"grad_norm": 0.08459590375423431,
|
| 241 |
+
"learning_rate": 4.893298743830168e-06,
|
| 242 |
+
"logits/chosen": 13.477182388305664,
|
| 243 |
+
"logits/rejected": 13.919464111328125,
|
| 244 |
+
"logps/chosen": -0.2495063841342926,
|
| 245 |
+
"logps/rejected": -0.3594816029071808,
|
| 246 |
+
"loss": 0.8957,
|
| 247 |
+
"rewards/accuracies": 0.612500011920929,
|
| 248 |
+
"rewards/chosen": -0.3742595613002777,
|
| 249 |
+
"rewards/margins": 0.16496284306049347,
|
| 250 |
+
"rewards/rejected": -0.5392224192619324,
|
| 251 |
+
"step": 140
|
| 252 |
+
},
|
| 253 |
+
{
|
| 254 |
+
"epoch": 0.12108980827447023,
|
| 255 |
+
"grad_norm": 0.12467797845602036,
|
| 256 |
+
"learning_rate": 4.8776412907378845e-06,
|
| 257 |
+
"logits/chosen": 12.801656723022461,
|
| 258 |
+
"logits/rejected": 13.564155578613281,
|
| 259 |
+
"logps/chosen": -0.2779986262321472,
|
| 260 |
+
"logps/rejected": -0.33566445112228394,
|
| 261 |
+
"loss": 0.892,
|
| 262 |
+
"rewards/accuracies": 0.4375,
|
| 263 |
+
"rewards/chosen": -0.4169979691505432,
|
| 264 |
+
"rewards/margins": 0.08649872243404388,
|
| 265 |
+
"rewards/rejected": -0.5034967064857483,
|
| 266 |
+
"step": 150
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 0.12108980827447023,
|
| 270 |
+
"eval_logits/chosen": 12.169118881225586,
|
| 271 |
+
"eval_logits/rejected": 13.019683837890625,
|
| 272 |
+
"eval_logps/chosen": -0.25547870993614197,
|
| 273 |
+
"eval_logps/rejected": -0.37277930974960327,
|
| 274 |
+
"eval_loss": 0.8789658546447754,
|
| 275 |
+
"eval_rewards/accuracies": 0.5742574334144592,
|
| 276 |
+
"eval_rewards/chosen": -0.38321802020072937,
|
| 277 |
+
"eval_rewards/margins": 0.17595094442367554,
|
| 278 |
+
"eval_rewards/rejected": -0.5591689944267273,
|
| 279 |
+
"eval_runtime": 29.0961,
|
| 280 |
+
"eval_samples_per_second": 27.529,
|
| 281 |
+
"eval_steps_per_second": 3.471,
|
| 282 |
+
"step": 150
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 0.12916246215943492,
|
| 286 |
+
"grad_norm": 0.1712370663881302,
|
| 287 |
+
"learning_rate": 4.860940925593703e-06,
|
| 288 |
+
"logits/chosen": 11.972528457641602,
|
| 289 |
+
"logits/rejected": 12.684088706970215,
|
| 290 |
+
"logps/chosen": -0.28340521454811096,
|
| 291 |
+
"logps/rejected": -0.3864063024520874,
|
| 292 |
+
"loss": 0.8789,
|
| 293 |
+
"rewards/accuracies": 0.574999988079071,
|
| 294 |
+
"rewards/chosen": -0.42510780692100525,
|
| 295 |
+
"rewards/margins": 0.15450166165828705,
|
| 296 |
+
"rewards/rejected": -0.5796095132827759,
|
| 297 |
+
"step": 160
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 0.13723511604439959,
|
| 301 |
+
"grad_norm": 0.21750673651695251,
|
| 302 |
+
"learning_rate": 4.84320497372973e-06,
|
| 303 |
+
"logits/chosen": 11.93881607055664,
|
| 304 |
+
"logits/rejected": 12.458230972290039,
|
| 305 |
+
"logps/chosen": -0.274599552154541,
|
| 306 |
+
"logps/rejected": -0.3882916271686554,
|
| 307 |
+
"loss": 0.8857,
|
| 308 |
+
"rewards/accuracies": 0.675000011920929,
|
| 309 |
+
"rewards/chosen": -0.4118993282318115,
|
| 310 |
+
"rewards/margins": 0.1705380380153656,
|
| 311 |
+
"rewards/rejected": -0.5824374556541443,
|
| 312 |
+
"step": 170
|
| 313 |
+
},
|
| 314 |
+
{
|
| 315 |
+
"epoch": 0.14530776992936428,
|
| 316 |
+
"grad_norm": 0.16000741720199585,
|
| 317 |
+
"learning_rate": 4.824441214720629e-06,
|
| 318 |
+
"logits/chosen": 9.631464958190918,
|
| 319 |
+
"logits/rejected": 10.772969245910645,
|
| 320 |
+
"logps/chosen": -0.287865549325943,
|
| 321 |
+
"logps/rejected": -0.4817379415035248,
|
| 322 |
+
"loss": 0.8699,
|
| 323 |
+
"rewards/accuracies": 0.6625000238418579,
|
| 324 |
+
"rewards/chosen": -0.4317983090877533,
|
| 325 |
+
"rewards/margins": 0.29080861806869507,
|
| 326 |
+
"rewards/rejected": -0.722606897354126,
|
| 327 |
+
"step": 180
|
| 328 |
+
},
|
| 329 |
+
{
|
| 330 |
+
"epoch": 0.15338042381432895,
|
| 331 |
+
"grad_norm": 0.46293890476226807,
|
| 332 |
+
"learning_rate": 4.804657878971252e-06,
|
| 333 |
+
"logits/chosen": 9.081937789916992,
|
| 334 |
+
"logits/rejected": 10.024572372436523,
|
| 335 |
+
"logps/chosen": -0.2941994071006775,
|
| 336 |
+
"logps/rejected": -0.4772109389305115,
|
| 337 |
+
"loss": 0.8565,
|
| 338 |
+
"rewards/accuracies": 0.675000011920929,
|
| 339 |
+
"rewards/chosen": -0.44129911065101624,
|
| 340 |
+
"rewards/margins": 0.2745172679424286,
|
| 341 |
+
"rewards/rejected": -0.7158163785934448,
|
| 342 |
+
"step": 190
|
| 343 |
+
},
|
| 344 |
+
{
|
| 345 |
+
"epoch": 0.16145307769929365,
|
| 346 |
+
"grad_norm": 0.24658434092998505,
|
| 347 |
+
"learning_rate": 4.783863644106502e-06,
|
| 348 |
+
"logits/chosen": 7.767125606536865,
|
| 349 |
+
"logits/rejected": 8.009145736694336,
|
| 350 |
+
"logps/chosen": -0.32877305150032043,
|
| 351 |
+
"logps/rejected": -0.47733697295188904,
|
| 352 |
+
"loss": 0.8482,
|
| 353 |
+
"rewards/accuracies": 0.5874999761581421,
|
| 354 |
+
"rewards/chosen": -0.49315959215164185,
|
| 355 |
+
"rewards/margins": 0.2228458821773529,
|
| 356 |
+
"rewards/rejected": -0.7160054445266724,
|
| 357 |
+
"step": 200
|
| 358 |
+
},
|
| 359 |
+
{
|
| 360 |
+
"epoch": 0.16145307769929365,
|
| 361 |
+
"eval_logits/chosen": 6.716187953948975,
|
| 362 |
+
"eval_logits/rejected": 7.248146057128906,
|
| 363 |
+
"eval_logps/chosen": -0.30652713775634766,
|
| 364 |
+
"eval_logps/rejected": -0.5056277513504028,
|
| 365 |
+
"eval_loss": 0.8243693113327026,
|
| 366 |
+
"eval_rewards/accuracies": 0.603960394859314,
|
| 367 |
+
"eval_rewards/chosen": -0.4597907066345215,
|
| 368 |
+
"eval_rewards/margins": 0.2986510097980499,
|
| 369 |
+
"eval_rewards/rejected": -0.7584417462348938,
|
| 370 |
+
"eval_runtime": 29.0989,
|
| 371 |
+
"eval_samples_per_second": 27.527,
|
| 372 |
+
"eval_steps_per_second": 3.471,
|
| 373 |
+
"step": 200
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.16952573158425832,
|
| 377 |
+
"grad_norm": 0.40796443819999695,
|
| 378 |
+
"learning_rate": 4.762067631165049e-06,
|
| 379 |
+
"logits/chosen": 5.932644844055176,
|
| 380 |
+
"logits/rejected": 6.521953582763672,
|
| 381 |
+
"logps/chosen": -0.32976508140563965,
|
| 382 |
+
"logps/rejected": -0.5628186464309692,
|
| 383 |
+
"loss": 0.7988,
|
| 384 |
+
"rewards/accuracies": 0.612500011920929,
|
| 385 |
+
"rewards/chosen": -0.4946475923061371,
|
| 386 |
+
"rewards/margins": 0.3495803475379944,
|
| 387 |
+
"rewards/rejected": -0.8442279696464539,
|
| 388 |
+
"step": 210
|
| 389 |
+
},
|
| 390 |
+
{
|
| 391 |
+
"epoch": 0.17759838546922302,
|
| 392 |
+
"grad_norm": 0.42468318343162537,
|
| 393 |
+
"learning_rate": 4.7392794005985324e-06,
|
| 394 |
+
"logits/chosen": 6.069305896759033,
|
| 395 |
+
"logits/rejected": 5.8950395584106445,
|
| 396 |
+
"logps/chosen": -0.37205421924591064,
|
| 397 |
+
"logps/rejected": -0.6190425753593445,
|
| 398 |
+
"loss": 0.796,
|
| 399 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 400 |
+
"rewards/chosen": -0.5580812692642212,
|
| 401 |
+
"rewards/margins": 0.3704826235771179,
|
| 402 |
+
"rewards/rejected": -0.9285639524459839,
|
| 403 |
+
"step": 220
|
| 404 |
+
},
|
| 405 |
+
{
|
| 406 |
+
"epoch": 0.1856710393541877,
|
| 407 |
+
"grad_norm": 0.4138280153274536,
|
| 408 |
+
"learning_rate": 4.715508948078037e-06,
|
| 409 |
+
"logits/chosen": 3.5271706581115723,
|
| 410 |
+
"logits/rejected": 3.363534927368164,
|
| 411 |
+
"logps/chosen": -0.41677650809288025,
|
| 412 |
+
"logps/rejected": -0.7121980786323547,
|
| 413 |
+
"loss": 0.7457,
|
| 414 |
+
"rewards/accuracies": 0.637499988079071,
|
| 415 |
+
"rewards/chosen": -0.6251648664474487,
|
| 416 |
+
"rewards/margins": 0.4431324005126953,
|
| 417 |
+
"rewards/rejected": -1.0682971477508545,
|
| 418 |
+
"step": 230
|
| 419 |
+
},
|
| 420 |
+
{
|
| 421 |
+
"epoch": 0.19374369323915236,
|
| 422 |
+
"grad_norm": 1.4327284097671509,
|
| 423 |
+
"learning_rate": 4.690766700109659e-06,
|
| 424 |
+
"logits/chosen": 2.296924114227295,
|
| 425 |
+
"logits/rejected": 1.6135867834091187,
|
| 426 |
+
"logps/chosen": -0.506227970123291,
|
| 427 |
+
"logps/rejected": -0.8080593943595886,
|
| 428 |
+
"loss": 0.7453,
|
| 429 |
+
"rewards/accuracies": 0.550000011920929,
|
| 430 |
+
"rewards/chosen": -0.7593418955802917,
|
| 431 |
+
"rewards/margins": 0.4527471661567688,
|
| 432 |
+
"rewards/rejected": -1.21208918094635,
|
| 433 |
+
"step": 240
|
| 434 |
+
},
|
| 435 |
+
{
|
| 436 |
+
"epoch": 0.20181634712411706,
|
| 437 |
+
"grad_norm": 0.4413074553012848,
|
| 438 |
+
"learning_rate": 4.665063509461098e-06,
|
| 439 |
+
"logits/chosen": 1.4450469017028809,
|
| 440 |
+
"logits/rejected": 0.40727120637893677,
|
| 441 |
+
"logps/chosen": -0.4749869406223297,
|
| 442 |
+
"logps/rejected": -0.9967275857925415,
|
| 443 |
+
"loss": 0.704,
|
| 444 |
+
"rewards/accuracies": 0.6625000238418579,
|
| 445 |
+
"rewards/chosen": -0.712480366230011,
|
| 446 |
+
"rewards/margins": 0.7826108932495117,
|
| 447 |
+
"rewards/rejected": -1.4950913190841675,
|
| 448 |
+
"step": 250
|
| 449 |
+
},
|
| 450 |
+
{
|
| 451 |
+
"epoch": 0.20181634712411706,
|
| 452 |
+
"eval_logits/chosen": 1.4614256620407104,
|
| 453 |
+
"eval_logits/rejected": 0.6895493865013123,
|
| 454 |
+
"eval_logps/chosen": -0.5038881301879883,
|
| 455 |
+
"eval_logps/rejected": -0.973581850528717,
|
| 456 |
+
"eval_loss": 0.6982013583183289,
|
| 457 |
+
"eval_rewards/accuracies": 0.6138613820075989,
|
| 458 |
+
"eval_rewards/chosen": -0.7558321952819824,
|
| 459 |
+
"eval_rewards/margins": 0.7045406699180603,
|
| 460 |
+
"eval_rewards/rejected": -1.460372805595398,
|
| 461 |
+
"eval_runtime": 29.0993,
|
| 462 |
+
"eval_samples_per_second": 27.526,
|
| 463 |
+
"eval_steps_per_second": 3.471,
|
| 464 |
+
"step": 250
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.20988900100908173,
|
| 468 |
+
"grad_norm": 0.734251856803894,
|
| 469 |
+
"learning_rate": 4.638410650401267e-06,
|
| 470 |
+
"logits/chosen": 2.4274039268493652,
|
| 471 |
+
"logits/rejected": 1.498230218887329,
|
| 472 |
+
"logps/chosen": -0.5549699068069458,
|
| 473 |
+
"logps/rejected": -0.9348627328872681,
|
| 474 |
+
"loss": 0.7149,
|
| 475 |
+
"rewards/accuracies": 0.5249999761581421,
|
| 476 |
+
"rewards/chosen": -0.8324548602104187,
|
| 477 |
+
"rewards/margins": 0.5698392987251282,
|
| 478 |
+
"rewards/rejected": -1.4022941589355469,
|
| 479 |
+
"step": 260
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 0.21796165489404642,
|
| 483 |
+
"grad_norm": 0.4445085823535919,
|
| 484 |
+
"learning_rate": 4.610819813755038e-06,
|
| 485 |
+
"logits/chosen": 2.042858600616455,
|
| 486 |
+
"logits/rejected": 1.1695036888122559,
|
| 487 |
+
"logps/chosen": -0.5573975443840027,
|
| 488 |
+
"logps/rejected": -1.15065598487854,
|
| 489 |
+
"loss": 0.6811,
|
| 490 |
+
"rewards/accuracies": 0.625,
|
| 491 |
+
"rewards/chosen": -0.8360962867736816,
|
| 492 |
+
"rewards/margins": 0.889887809753418,
|
| 493 |
+
"rewards/rejected": -1.72598397731781,
|
| 494 |
+
"step": 270
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 0.2260343087790111,
|
| 498 |
+
"grad_norm": 0.5875476598739624,
|
| 499 |
+
"learning_rate": 4.582303101775249e-06,
|
| 500 |
+
"logits/chosen": 1.3029518127441406,
|
| 501 |
+
"logits/rejected": 0.3808741271495819,
|
| 502 |
+
"logps/chosen": -0.6656169891357422,
|
| 503 |
+
"logps/rejected": -1.5305824279785156,
|
| 504 |
+
"loss": 0.637,
|
| 505 |
+
"rewards/accuracies": 0.625,
|
| 506 |
+
"rewards/chosen": -0.9984253644943237,
|
| 507 |
+
"rewards/margins": 1.2974482774734497,
|
| 508 |
+
"rewards/rejected": -2.2958736419677734,
|
| 509 |
+
"step": 280
|
| 510 |
+
},
|
| 511 |
+
{
|
| 512 |
+
"epoch": 0.2341069626639758,
|
| 513 |
+
"grad_norm": 0.3814420998096466,
|
| 514 |
+
"learning_rate": 4.55287302283426e-06,
|
| 515 |
+
"logits/chosen": 2.0200212001800537,
|
| 516 |
+
"logits/rejected": 1.2681838274002075,
|
| 517 |
+
"logps/chosen": -0.6776013970375061,
|
| 518 |
+
"logps/rejected": -1.3369777202606201,
|
| 519 |
+
"loss": 0.63,
|
| 520 |
+
"rewards/accuracies": 0.5874999761581421,
|
| 521 |
+
"rewards/chosen": -1.016402006149292,
|
| 522 |
+
"rewards/margins": 0.9890643358230591,
|
| 523 |
+
"rewards/rejected": -2.0054664611816406,
|
| 524 |
+
"step": 290
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 0.24217961654894046,
|
| 528 |
+
"grad_norm": 0.6442322731018066,
|
| 529 |
+
"learning_rate": 4.522542485937369e-06,
|
| 530 |
+
"logits/chosen": 1.6268196105957031,
|
| 531 |
+
"logits/rejected": 0.4954712390899658,
|
| 532 |
+
"logps/chosen": -0.8117648363113403,
|
| 533 |
+
"logps/rejected": -1.7705228328704834,
|
| 534 |
+
"loss": 0.6104,
|
| 535 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 536 |
+
"rewards/chosen": -1.2176473140716553,
|
| 537 |
+
"rewards/margins": 1.4381370544433594,
|
| 538 |
+
"rewards/rejected": -2.6557843685150146,
|
| 539 |
+
"step": 300
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"epoch": 0.24217961654894046,
|
| 543 |
+
"eval_logits/chosen": 1.4003607034683228,
|
| 544 |
+
"eval_logits/rejected": 0.46628010272979736,
|
| 545 |
+
"eval_logps/chosen": -0.7612115740776062,
|
| 546 |
+
"eval_logps/rejected": -1.6895866394042969,
|
| 547 |
+
"eval_loss": 0.5781419277191162,
|
| 548 |
+
"eval_rewards/accuracies": 0.6534653306007385,
|
| 549 |
+
"eval_rewards/chosen": -1.141817331314087,
|
| 550 |
+
"eval_rewards/margins": 1.392562747001648,
|
| 551 |
+
"eval_rewards/rejected": -2.5343799591064453,
|
| 552 |
+
"eval_runtime": 29.102,
|
| 553 |
+
"eval_samples_per_second": 27.524,
|
| 554 |
+
"eval_steps_per_second": 3.471,
|
| 555 |
+
"step": 300
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.25025227043390513,
|
| 559 |
+
"grad_norm": 0.7889758944511414,
|
| 560 |
+
"learning_rate": 4.491324795060491e-06,
|
| 561 |
+
"logits/chosen": 1.1702089309692383,
|
| 562 |
+
"logits/rejected": 0.354276180267334,
|
| 563 |
+
"logps/chosen": -0.8091352581977844,
|
| 564 |
+
"logps/rejected": -1.944819450378418,
|
| 565 |
+
"loss": 0.5807,
|
| 566 |
+
"rewards/accuracies": 0.6875,
|
| 567 |
+
"rewards/chosen": -1.2137027978897095,
|
| 568 |
+
"rewards/margins": 1.7035261392593384,
|
| 569 |
+
"rewards/rejected": -2.9172286987304688,
|
| 570 |
+
"step": 310
|
| 571 |
+
},
|
| 572 |
+
{
|
| 573 |
+
"epoch": 0.25832492431886983,
|
| 574 |
+
"grad_norm": 1.410145878791809,
|
| 575 |
+
"learning_rate": 4.4592336433146e-06,
|
| 576 |
+
"logits/chosen": 2.0981459617614746,
|
| 577 |
+
"logits/rejected": 1.2289329767227173,
|
| 578 |
+
"logps/chosen": -0.8151634931564331,
|
| 579 |
+
"logps/rejected": -1.977259874343872,
|
| 580 |
+
"loss": 0.5007,
|
| 581 |
+
"rewards/accuracies": 0.6625000238418579,
|
| 582 |
+
"rewards/chosen": -1.222745418548584,
|
| 583 |
+
"rewards/margins": 1.7431443929672241,
|
| 584 |
+
"rewards/rejected": -2.9658896923065186,
|
| 585 |
+
"step": 320
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 0.26639757820383453,
|
| 589 |
+
"grad_norm": 0.5564689040184021,
|
| 590 |
+
"learning_rate": 4.426283106939474e-06,
|
| 591 |
+
"logits/chosen": 1.7921768426895142,
|
| 592 |
+
"logits/rejected": 0.7705962061882019,
|
| 593 |
+
"logps/chosen": -0.9244564771652222,
|
| 594 |
+
"logps/rejected": -2.5274672508239746,
|
| 595 |
+
"loss": 0.4757,
|
| 596 |
+
"rewards/accuracies": 0.75,
|
| 597 |
+
"rewards/chosen": -1.3866846561431885,
|
| 598 |
+
"rewards/margins": 2.4045166969299316,
|
| 599 |
+
"rewards/rejected": -3.791201114654541,
|
| 600 |
+
"step": 330
|
| 601 |
+
},
|
| 602 |
+
{
|
| 603 |
+
"epoch": 0.27447023208879917,
|
| 604 |
+
"grad_norm": 0.7554243803024292,
|
| 605 |
+
"learning_rate": 4.3924876391293915e-06,
|
| 606 |
+
"logits/chosen": 1.947997808456421,
|
| 607 |
+
"logits/rejected": 0.9661592245101929,
|
| 608 |
+
"logps/chosen": -0.9154227375984192,
|
| 609 |
+
"logps/rejected": -3.0766491889953613,
|
| 610 |
+
"loss": 0.4581,
|
| 611 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 612 |
+
"rewards/chosen": -1.3731342554092407,
|
| 613 |
+
"rewards/margins": 3.24183988571167,
|
| 614 |
+
"rewards/rejected": -4.614974021911621,
|
| 615 |
+
"step": 340
|
| 616 |
+
},
|
| 617 |
+
{
|
| 618 |
+
"epoch": 0.28254288597376387,
|
| 619 |
+
"grad_norm": 2.936426877975464,
|
| 620 |
+
"learning_rate": 4.357862063693486e-06,
|
| 621 |
+
"logits/chosen": 1.2846364974975586,
|
| 622 |
+
"logits/rejected": 0.9968118667602539,
|
| 623 |
+
"logps/chosen": -1.0272481441497803,
|
| 624 |
+
"logps/rejected": -2.6980624198913574,
|
| 625 |
+
"loss": 0.4989,
|
| 626 |
+
"rewards/accuracies": 0.7124999761581421,
|
| 627 |
+
"rewards/chosen": -1.54087233543396,
|
| 628 |
+
"rewards/margins": 2.506221294403076,
|
| 629 |
+
"rewards/rejected": -4.047093868255615,
|
| 630 |
+
"step": 350
|
| 631 |
+
},
|
| 632 |
+
{
|
| 633 |
+
"epoch": 0.28254288597376387,
|
| 634 |
+
"eval_logits/chosen": 1.841333031654358,
|
| 635 |
+
"eval_logits/rejected": 1.1651691198349,
|
| 636 |
+
"eval_logps/chosen": -0.9644113183021545,
|
| 637 |
+
"eval_logps/rejected": -2.631535053253174,
|
| 638 |
+
"eval_loss": 0.44154587388038635,
|
| 639 |
+
"eval_rewards/accuracies": 0.6732673048973083,
|
| 640 |
+
"eval_rewards/chosen": -1.4466170072555542,
|
| 641 |
+
"eval_rewards/margins": 2.500684976577759,
|
| 642 |
+
"eval_rewards/rejected": -3.9473025798797607,
|
| 643 |
+
"eval_runtime": 29.1062,
|
| 644 |
+
"eval_samples_per_second": 27.52,
|
| 645 |
+
"eval_steps_per_second": 3.47,
|
| 646 |
+
"step": 350
|
| 647 |
+
},
|
| 648 |
+
{
|
| 649 |
+
"epoch": 0.29061553985872857,
|
| 650 |
+
"grad_norm": 0.49919044971466064,
|
| 651 |
+
"learning_rate": 4.322421568553529e-06,
|
| 652 |
+
"logits/chosen": 1.9874728918075562,
|
| 653 |
+
"logits/rejected": 1.464611291885376,
|
| 654 |
+
"logps/chosen": -0.9963932037353516,
|
| 655 |
+
"logps/rejected": -2.8804218769073486,
|
| 656 |
+
"loss": 0.4373,
|
| 657 |
+
"rewards/accuracies": 0.7124999761581421,
|
| 658 |
+
"rewards/chosen": -1.4945898056030273,
|
| 659 |
+
"rewards/margins": 2.8260433673858643,
|
| 660 |
+
"rewards/rejected": -4.320633411407471,
|
| 661 |
+
"step": 360
|
| 662 |
+
},
|
| 663 |
+
{
|
| 664 |
+
"epoch": 0.29868819374369326,
|
| 665 |
+
"grad_norm": 1.4991668462753296,
|
| 666 |
+
"learning_rate": 4.286181699082008e-06,
|
| 667 |
+
"logits/chosen": 2.5421957969665527,
|
| 668 |
+
"logits/rejected": 1.941922903060913,
|
| 669 |
+
"logps/chosen": -1.0214306116104126,
|
| 670 |
+
"logps/rejected": -3.2893283367156982,
|
| 671 |
+
"loss": 0.438,
|
| 672 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 673 |
+
"rewards/chosen": -1.5321458578109741,
|
| 674 |
+
"rewards/margins": 3.4018466472625732,
|
| 675 |
+
"rewards/rejected": -4.933992862701416,
|
| 676 |
+
"step": 370
|
| 677 |
+
},
|
| 678 |
+
{
|
| 679 |
+
"epoch": 0.3067608476286579,
|
| 680 |
+
"grad_norm": 1.933100938796997,
|
| 681 |
+
"learning_rate": 4.249158351283414e-06,
|
| 682 |
+
"logits/chosen": 1.80439031124115,
|
| 683 |
+
"logits/rejected": 1.4421275854110718,
|
| 684 |
+
"logps/chosen": -1.1270357370376587,
|
| 685 |
+
"logps/rejected": -3.17500901222229,
|
| 686 |
+
"loss": 0.4542,
|
| 687 |
+
"rewards/accuracies": 0.7124999761581421,
|
| 688 |
+
"rewards/chosen": -1.6905533075332642,
|
| 689 |
+
"rewards/margins": 3.071959972381592,
|
| 690 |
+
"rewards/rejected": -4.762513160705566,
|
| 691 |
+
"step": 380
|
| 692 |
+
},
|
| 693 |
+
{
|
| 694 |
+
"epoch": 0.3148335015136226,
|
| 695 |
+
"grad_norm": 3.017254590988159,
|
| 696 |
+
"learning_rate": 4.211367764821722e-06,
|
| 697 |
+
"logits/chosen": 3.4090209007263184,
|
| 698 |
+
"logits/rejected": 2.784639835357666,
|
| 699 |
+
"logps/chosen": -1.1713608503341675,
|
| 700 |
+
"logps/rejected": -3.326244831085205,
|
| 701 |
+
"loss": 0.4816,
|
| 702 |
+
"rewards/accuracies": 0.675000011920929,
|
| 703 |
+
"rewards/chosen": -1.757041335105896,
|
| 704 |
+
"rewards/margins": 3.232325792312622,
|
| 705 |
+
"rewards/rejected": -4.9893670082092285,
|
| 706 |
+
"step": 390
|
| 707 |
+
},
|
| 708 |
+
{
|
| 709 |
+
"epoch": 0.3229061553985873,
|
| 710 |
+
"grad_norm": 0.5897337198257446,
|
| 711 |
+
"learning_rate": 4.172826515897146e-06,
|
| 712 |
+
"logits/chosen": 3.6071503162384033,
|
| 713 |
+
"logits/rejected": 2.738395929336548,
|
| 714 |
+
"logps/chosen": -1.133429765701294,
|
| 715 |
+
"logps/rejected": -3.509474277496338,
|
| 716 |
+
"loss": 0.4431,
|
| 717 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 718 |
+
"rewards/chosen": -1.7001447677612305,
|
| 719 |
+
"rewards/margins": 3.5640671253204346,
|
| 720 |
+
"rewards/rejected": -5.264212131500244,
|
| 721 |
+
"step": 400
|
| 722 |
+
},
|
| 723 |
+
{
|
| 724 |
+
"epoch": 0.3229061553985873,
|
| 725 |
+
"eval_logits/chosen": 2.893691062927246,
|
| 726 |
+
"eval_logits/rejected": 2.259718656539917,
|
| 727 |
+
"eval_logps/chosen": -1.0705492496490479,
|
| 728 |
+
"eval_logps/rejected": -2.935904026031494,
|
| 729 |
+
"eval_loss": 0.4156961143016815,
|
| 730 |
+
"eval_rewards/accuracies": 0.6732673048973083,
|
| 731 |
+
"eval_rewards/chosen": -1.6058237552642822,
|
| 732 |
+
"eval_rewards/margins": 2.798032283782959,
|
| 733 |
+
"eval_rewards/rejected": -4.403855800628662,
|
| 734 |
+
"eval_runtime": 29.186,
|
| 735 |
+
"eval_samples_per_second": 27.445,
|
| 736 |
+
"eval_steps_per_second": 3.461,
|
| 737 |
+
"step": 400
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 0.33097880928355194,
|
| 741 |
+
"grad_norm": 2.219709634780884,
|
| 742 |
+
"learning_rate": 4.133551509975264e-06,
|
| 743 |
+
"logits/chosen": 2.8940415382385254,
|
| 744 |
+
"logits/rejected": 2.545257091522217,
|
| 745 |
+
"logps/chosen": -1.1774877309799194,
|
| 746 |
+
"logps/rejected": -3.4022269248962402,
|
| 747 |
+
"loss": 0.427,
|
| 748 |
+
"rewards/accuracies": 0.699999988079071,
|
| 749 |
+
"rewards/chosen": -1.7662317752838135,
|
| 750 |
+
"rewards/margins": 3.3371081352233887,
|
| 751 |
+
"rewards/rejected": -5.1033406257629395,
|
| 752 |
+
"step": 410
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 0.33905146316851664,
|
| 756 |
+
"grad_norm": 0.9891649484634399,
|
| 757 |
+
"learning_rate": 4.093559974371725e-06,
|
| 758 |
+
"logits/chosen": 3.180427074432373,
|
| 759 |
+
"logits/rejected": 2.8354597091674805,
|
| 760 |
+
"logps/chosen": -1.2984545230865479,
|
| 761 |
+
"logps/rejected": -3.4927687644958496,
|
| 762 |
+
"loss": 0.4384,
|
| 763 |
+
"rewards/accuracies": 0.75,
|
| 764 |
+
"rewards/chosen": -1.9476817846298218,
|
| 765 |
+
"rewards/margins": 3.291471481323242,
|
| 766 |
+
"rewards/rejected": -5.2391533851623535,
|
| 767 |
+
"step": 420
|
| 768 |
+
},
|
| 769 |
+
{
|
| 770 |
+
"epoch": 0.34712411705348134,
|
| 771 |
+
"grad_norm": 0.4516288936138153,
|
| 772 |
+
"learning_rate": 4.052869450695776e-06,
|
| 773 |
+
"logits/chosen": 4.026360511779785,
|
| 774 |
+
"logits/rejected": 3.3932456970214844,
|
| 775 |
+
"logps/chosen": -1.1195746660232544,
|
| 776 |
+
"logps/rejected": -3.4704582691192627,
|
| 777 |
+
"loss": 0.3934,
|
| 778 |
+
"rewards/accuracies": 0.6625000238418579,
|
| 779 |
+
"rewards/chosen": -1.6793619394302368,
|
| 780 |
+
"rewards/margins": 3.5263259410858154,
|
| 781 |
+
"rewards/rejected": -5.205687522888184,
|
| 782 |
+
"step": 430
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 0.35519677093844604,
|
| 786 |
+
"grad_norm": 0.5276560187339783,
|
| 787 |
+
"learning_rate": 4.011497787155938e-06,
|
| 788 |
+
"logits/chosen": 3.410902500152588,
|
| 789 |
+
"logits/rejected": 2.906996965408325,
|
| 790 |
+
"logps/chosen": -1.205904245376587,
|
| 791 |
+
"logps/rejected": -3.7434115409851074,
|
| 792 |
+
"loss": 0.4147,
|
| 793 |
+
"rewards/accuracies": 0.737500011920929,
|
| 794 |
+
"rewards/chosen": -1.8088566064834595,
|
| 795 |
+
"rewards/margins": 3.806260585784912,
|
| 796 |
+
"rewards/rejected": -5.615117073059082,
|
| 797 |
+
"step": 440
|
| 798 |
+
},
|
| 799 |
+
{
|
| 800 |
+
"epoch": 0.3632694248234107,
|
| 801 |
+
"grad_norm": 0.4435446560382843,
|
| 802 |
+
"learning_rate": 3.969463130731183e-06,
|
| 803 |
+
"logits/chosen": 3.7616615295410156,
|
| 804 |
+
"logits/rejected": 3.4183363914489746,
|
| 805 |
+
"logps/chosen": -1.2180852890014648,
|
| 806 |
+
"logps/rejected": -3.865541934967041,
|
| 807 |
+
"loss": 0.4354,
|
| 808 |
+
"rewards/accuracies": 0.675000011920929,
|
| 809 |
+
"rewards/chosen": -1.8271278142929077,
|
| 810 |
+
"rewards/margins": 3.9711856842041016,
|
| 811 |
+
"rewards/rejected": -5.798312664031982,
|
| 812 |
+
"step": 450
|
| 813 |
+
},
|
| 814 |
+
{
|
| 815 |
+
"epoch": 0.3632694248234107,
|
| 816 |
+
"eval_logits/chosen": 3.55381178855896,
|
| 817 |
+
"eval_logits/rejected": 3.01943039894104,
|
| 818 |
+
"eval_logps/chosen": -1.3239827156066895,
|
| 819 |
+
"eval_logps/rejected": -3.6477084159851074,
|
| 820 |
+
"eval_loss": 0.3905698359012604,
|
| 821 |
+
"eval_rewards/accuracies": 0.7029703259468079,
|
| 822 |
+
"eval_rewards/chosen": -1.9859741926193237,
|
| 823 |
+
"eval_rewards/margins": 3.485589027404785,
|
| 824 |
+
"eval_rewards/rejected": -5.471563816070557,
|
| 825 |
+
"eval_runtime": 29.0958,
|
| 826 |
+
"eval_samples_per_second": 27.53,
|
| 827 |
+
"eval_steps_per_second": 3.471,
|
| 828 |
+
"step": 450
|
| 829 |
+
},
|
| 830 |
+
{
|
| 831 |
+
"epoch": 0.3713420787083754,
|
| 832 |
+
"grad_norm": 0.8419796824455261,
|
| 833 |
+
"learning_rate": 3.92678391921108e-06,
|
| 834 |
+
"logits/chosen": 3.3496861457824707,
|
| 835 |
+
"logits/rejected": 3.235544204711914,
|
| 836 |
+
"logps/chosen": -1.4899990558624268,
|
| 837 |
+
"logps/rejected": -4.507351398468018,
|
| 838 |
+
"loss": 0.4023,
|
| 839 |
+
"rewards/accuracies": 0.762499988079071,
|
| 840 |
+
"rewards/chosen": -2.2349984645843506,
|
| 841 |
+
"rewards/margins": 4.526029109954834,
|
| 842 |
+
"rewards/rejected": -6.7610273361206055,
|
| 843 |
+
"step": 460
|
| 844 |
+
},
|
| 845 |
+
{
|
| 846 |
+
"epoch": 0.3794147325933401,
|
| 847 |
+
"grad_norm": 1.2379052639007568,
|
| 848 |
+
"learning_rate": 3.88347887310836e-06,
|
| 849 |
+
"logits/chosen": 3.5375685691833496,
|
| 850 |
+
"logits/rejected": 3.34586763381958,
|
| 851 |
+
"logps/chosen": -1.5865886211395264,
|
| 852 |
+
"logps/rejected": -4.887473106384277,
|
| 853 |
+
"loss": 0.3979,
|
| 854 |
+
"rewards/accuracies": 0.7875000238418579,
|
| 855 |
+
"rewards/chosen": -2.379883289337158,
|
| 856 |
+
"rewards/margins": 4.9513258934021,
|
| 857 |
+
"rewards/rejected": -7.331210136413574,
|
| 858 |
+
"step": 470
|
| 859 |
+
},
|
| 860 |
+
{
|
| 861 |
+
"epoch": 0.3874873864783047,
|
| 862 |
+
"grad_norm": 1.9969083070755005,
|
| 863 |
+
"learning_rate": 3.839566987447492e-06,
|
| 864 |
+
"logits/chosen": 3.3116583824157715,
|
| 865 |
+
"logits/rejected": 2.9652373790740967,
|
| 866 |
+
"logps/chosen": -2.0900139808654785,
|
| 867 |
+
"logps/rejected": -4.921124458312988,
|
| 868 |
+
"loss": 0.3356,
|
| 869 |
+
"rewards/accuracies": 0.824999988079071,
|
| 870 |
+
"rewards/chosen": -3.135021209716797,
|
| 871 |
+
"rewards/margins": 4.246665000915527,
|
| 872 |
+
"rewards/rejected": -7.381686210632324,
|
| 873 |
+
"step": 480
|
| 874 |
+
},
|
| 875 |
+
{
|
| 876 |
+
"epoch": 0.3955600403632694,
|
| 877 |
+
"grad_norm": 2.3405466079711914,
|
| 878 |
+
"learning_rate": 3.795067523432826e-06,
|
| 879 |
+
"logits/chosen": 2.542064666748047,
|
| 880 |
+
"logits/rejected": 2.503020763397217,
|
| 881 |
+
"logps/chosen": -2.4674277305603027,
|
| 882 |
+
"logps/rejected": -5.703408718109131,
|
| 883 |
+
"loss": 0.3193,
|
| 884 |
+
"rewards/accuracies": 0.875,
|
| 885 |
+
"rewards/chosen": -3.701141834259033,
|
| 886 |
+
"rewards/margins": 4.853971004486084,
|
| 887 |
+
"rewards/rejected": -8.555112838745117,
|
| 888 |
+
"step": 490
|
| 889 |
+
},
|
| 890 |
+
{
|
| 891 |
+
"epoch": 0.4036326942482341,
|
| 892 |
+
"grad_norm": 1.4798164367675781,
|
| 893 |
+
"learning_rate": 3.7500000000000005e-06,
|
| 894 |
+
"logits/chosen": 3.79949688911438,
|
| 895 |
+
"logits/rejected": 3.7085328102111816,
|
| 896 |
+
"logps/chosen": -2.6355605125427246,
|
| 897 |
+
"logps/rejected": -5.005209922790527,
|
| 898 |
+
"loss": 0.3103,
|
| 899 |
+
"rewards/accuracies": 0.8125,
|
| 900 |
+
"rewards/chosen": -3.953340530395508,
|
| 901 |
+
"rewards/margins": 3.554474353790283,
|
| 902 |
+
"rewards/rejected": -7.507814884185791,
|
| 903 |
+
"step": 500
|
| 904 |
+
},
|
| 905 |
+
{
|
| 906 |
+
"epoch": 0.4036326942482341,
|
| 907 |
+
"eval_logits/chosen": 3.31948184967041,
|
| 908 |
+
"eval_logits/rejected": 2.9104762077331543,
|
| 909 |
+
"eval_logps/chosen": -2.606774091720581,
|
| 910 |
+
"eval_logps/rejected": -5.2280120849609375,
|
| 911 |
+
"eval_loss": 0.3228474259376526,
|
| 912 |
+
"eval_rewards/accuracies": 0.9306930899620056,
|
| 913 |
+
"eval_rewards/chosen": -3.9101614952087402,
|
| 914 |
+
"eval_rewards/margins": 3.931856870651245,
|
| 915 |
+
"eval_rewards/rejected": -7.8420186042785645,
|
| 916 |
+
"eval_runtime": 29.0975,
|
| 917 |
+
"eval_samples_per_second": 27.528,
|
| 918 |
+
"eval_steps_per_second": 3.471,
|
| 919 |
+
"step": 500
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"epoch": 0.4117053481331988,
|
| 923 |
+
"grad_norm": 1.4098985195159912,
|
| 924 |
+
"learning_rate": 3.7043841852542884e-06,
|
| 925 |
+
"logits/chosen": 2.69830584526062,
|
| 926 |
+
"logits/rejected": 2.7101688385009766,
|
| 927 |
+
"logps/chosen": -2.7831828594207764,
|
| 928 |
+
"logps/rejected": -5.963825225830078,
|
| 929 |
+
"loss": 0.2884,
|
| 930 |
+
"rewards/accuracies": 0.887499988079071,
|
| 931 |
+
"rewards/chosen": -4.174774169921875,
|
| 932 |
+
"rewards/margins": 4.7709641456604,
|
| 933 |
+
"rewards/rejected": -8.945737838745117,
|
| 934 |
+
"step": 510
|
| 935 |
+
},
|
| 936 |
+
{
|
| 937 |
+
"epoch": 0.41977800201816345,
|
| 938 |
+
"grad_norm": 3.457231283187866,
|
| 939 |
+
"learning_rate": 3.658240087799655e-06,
|
| 940 |
+
"logits/chosen": 3.335204601287842,
|
| 941 |
+
"logits/rejected": 3.1935055255889893,
|
| 942 |
+
"logps/chosen": -2.66495943069458,
|
| 943 |
+
"logps/rejected": -5.2761640548706055,
|
| 944 |
+
"loss": 0.3093,
|
| 945 |
+
"rewards/accuracies": 0.875,
|
| 946 |
+
"rewards/chosen": -3.997438907623291,
|
| 947 |
+
"rewards/margins": 3.916806697845459,
|
| 948 |
+
"rewards/rejected": -7.91424560546875,
|
| 949 |
+
"step": 520
|
| 950 |
+
},
|
| 951 |
+
{
|
| 952 |
+
"epoch": 0.42785065590312815,
|
| 953 |
+
"grad_norm": 1.7744654417037964,
|
| 954 |
+
"learning_rate": 3.611587947962319e-06,
|
| 955 |
+
"logits/chosen": 2.725924015045166,
|
| 956 |
+
"logits/rejected": 2.605053186416626,
|
| 957 |
+
"logps/chosen": -2.5286645889282227,
|
| 958 |
+
"logps/rejected": -5.774144172668457,
|
| 959 |
+
"loss": 0.2972,
|
| 960 |
+
"rewards/accuracies": 0.887499988079071,
|
| 961 |
+
"rewards/chosen": -3.792996883392334,
|
| 962 |
+
"rewards/margins": 4.868220329284668,
|
| 963 |
+
"rewards/rejected": -8.661216735839844,
|
| 964 |
+
"step": 530
|
| 965 |
+
},
|
| 966 |
+
{
|
| 967 |
+
"epoch": 0.43592330978809285,
|
| 968 |
+
"grad_norm": 2.863692283630371,
|
| 969 |
+
"learning_rate": 3.564448228912682e-06,
|
| 970 |
+
"logits/chosen": 2.304443120956421,
|
| 971 |
+
"logits/rejected": 2.4836020469665527,
|
| 972 |
+
"logps/chosen": -2.699148654937744,
|
| 973 |
+
"logps/rejected": -6.340291500091553,
|
| 974 |
+
"loss": 0.3136,
|
| 975 |
+
"rewards/accuracies": 0.887499988079071,
|
| 976 |
+
"rewards/chosen": -4.048723220825195,
|
| 977 |
+
"rewards/margins": 5.461714267730713,
|
| 978 |
+
"rewards/rejected": -9.510437965393066,
|
| 979 |
+
"step": 540
|
| 980 |
+
},
|
| 981 |
+
{
|
| 982 |
+
"epoch": 0.4439959636730575,
|
| 983 |
+
"grad_norm": 4.6681623458862305,
|
| 984 |
+
"learning_rate": 3.516841607689501e-06,
|
| 985 |
+
"logits/chosen": 3.1207594871520996,
|
| 986 |
+
"logits/rejected": 2.947653293609619,
|
| 987 |
+
"logps/chosen": -3.0062317848205566,
|
| 988 |
+
"logps/rejected": -6.26816463470459,
|
| 989 |
+
"loss": 0.2822,
|
| 990 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 991 |
+
"rewards/chosen": -4.509347438812256,
|
| 992 |
+
"rewards/margins": 4.892899513244629,
|
| 993 |
+
"rewards/rejected": -9.402246475219727,
|
| 994 |
+
"step": 550
|
| 995 |
+
},
|
| 996 |
+
{
|
| 997 |
+
"epoch": 0.4439959636730575,
|
| 998 |
+
"eval_logits/chosen": 2.879721164703369,
|
| 999 |
+
"eval_logits/rejected": 2.52551007270813,
|
| 1000 |
+
"eval_logps/chosen": -2.6468751430511475,
|
| 1001 |
+
"eval_logps/rejected": -5.636789321899414,
|
| 1002 |
+
"eval_loss": 0.2939932942390442,
|
| 1003 |
+
"eval_rewards/accuracies": 0.9306930899620056,
|
| 1004 |
+
"eval_rewards/chosen": -3.9703128337860107,
|
| 1005 |
+
"eval_rewards/margins": 4.4848713874816895,
|
| 1006 |
+
"eval_rewards/rejected": -8.455184936523438,
|
| 1007 |
+
"eval_runtime": 29.4025,
|
| 1008 |
+
"eval_samples_per_second": 27.243,
|
| 1009 |
+
"eval_steps_per_second": 3.435,
|
| 1010 |
+
"step": 550
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 0.4520686175580222,
|
| 1014 |
+
"grad_norm": 1.9975569248199463,
|
| 1015 |
+
"learning_rate": 3.4687889661302577e-06,
|
| 1016 |
+
"logits/chosen": 2.5885415077209473,
|
| 1017 |
+
"logits/rejected": 2.5374302864074707,
|
| 1018 |
+
"logps/chosen": -2.886643171310425,
|
| 1019 |
+
"logps/rejected": -6.636415004730225,
|
| 1020 |
+
"loss": 0.2736,
|
| 1021 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1022 |
+
"rewards/chosen": -4.329964637756348,
|
| 1023 |
+
"rewards/margins": 5.624656677246094,
|
| 1024 |
+
"rewards/rejected": -9.954621315002441,
|
| 1025 |
+
"step": 560
|
| 1026 |
+
},
|
| 1027 |
+
{
|
| 1028 |
+
"epoch": 0.4601412714429869,
|
| 1029 |
+
"grad_norm": 3.3965606689453125,
|
| 1030 |
+
"learning_rate": 3.4203113817116955e-06,
|
| 1031 |
+
"logits/chosen": 2.0093374252319336,
|
| 1032 |
+
"logits/rejected": 2.1064200401306152,
|
| 1033 |
+
"logps/chosen": -3.251981735229492,
|
| 1034 |
+
"logps/rejected": -6.749331474304199,
|
| 1035 |
+
"loss": 0.2776,
|
| 1036 |
+
"rewards/accuracies": 0.875,
|
| 1037 |
+
"rewards/chosen": -4.877972602844238,
|
| 1038 |
+
"rewards/margins": 5.246025085449219,
|
| 1039 |
+
"rewards/rejected": -10.12399673461914,
|
| 1040 |
+
"step": 570
|
| 1041 |
+
},
|
| 1042 |
+
{
|
| 1043 |
+
"epoch": 0.4682139253279516,
|
| 1044 |
+
"grad_norm": 2.7079927921295166,
|
| 1045 |
+
"learning_rate": 3.3714301183045382e-06,
|
| 1046 |
+
"logits/chosen": 3.916841506958008,
|
| 1047 |
+
"logits/rejected": 3.524440288543701,
|
| 1048 |
+
"logps/chosen": -3.304469347000122,
|
| 1049 |
+
"logps/rejected": -7.1993279457092285,
|
| 1050 |
+
"loss": 0.2547,
|
| 1051 |
+
"rewards/accuracies": 0.9375,
|
| 1052 |
+
"rewards/chosen": -4.956704139709473,
|
| 1053 |
+
"rewards/margins": 5.842287540435791,
|
| 1054 |
+
"rewards/rejected": -10.798992156982422,
|
| 1055 |
+
"step": 580
|
| 1056 |
+
},
|
| 1057 |
+
{
|
| 1058 |
+
"epoch": 0.47628657921291623,
|
| 1059 |
+
"grad_norm": 7.504203796386719,
|
| 1060 |
+
"learning_rate": 3.3221666168464584e-06,
|
| 1061 |
+
"logits/chosen": 2.7350666522979736,
|
| 1062 |
+
"logits/rejected": 2.2590534687042236,
|
| 1063 |
+
"logps/chosen": -3.0954322814941406,
|
| 1064 |
+
"logps/rejected": -6.782465934753418,
|
| 1065 |
+
"loss": 0.2839,
|
| 1066 |
+
"rewards/accuracies": 0.887499988079071,
|
| 1067 |
+
"rewards/chosen": -4.643147945404053,
|
| 1068 |
+
"rewards/margins": 5.530551433563232,
|
| 1069 |
+
"rewards/rejected": -10.173700332641602,
|
| 1070 |
+
"step": 590
|
| 1071 |
+
},
|
| 1072 |
+
{
|
| 1073 |
+
"epoch": 0.4843592330978809,
|
| 1074 |
+
"grad_norm": 2.8209569454193115,
|
| 1075 |
+
"learning_rate": 3.272542485937369e-06,
|
| 1076 |
+
"logits/chosen": 3.7743606567382812,
|
| 1077 |
+
"logits/rejected": 3.2480530738830566,
|
| 1078 |
+
"logps/chosen": -3.196892261505127,
|
| 1079 |
+
"logps/rejected": -6.42517614364624,
|
| 1080 |
+
"loss": 0.3237,
|
| 1081 |
+
"rewards/accuracies": 0.875,
|
| 1082 |
+
"rewards/chosen": -4.7953386306762695,
|
| 1083 |
+
"rewards/margins": 4.8424248695373535,
|
| 1084 |
+
"rewards/rejected": -9.637763977050781,
|
| 1085 |
+
"step": 600
|
| 1086 |
+
},
|
| 1087 |
+
{
|
| 1088 |
+
"epoch": 0.4843592330978809,
|
| 1089 |
+
"eval_logits/chosen": 2.6344597339630127,
|
| 1090 |
+
"eval_logits/rejected": 2.46531081199646,
|
| 1091 |
+
"eval_logps/chosen": -2.9877941608428955,
|
| 1092 |
+
"eval_logps/rejected": -6.266376972198486,
|
| 1093 |
+
"eval_loss": 0.2827170193195343,
|
| 1094 |
+
"eval_rewards/accuracies": 0.9603960514068604,
|
| 1095 |
+
"eval_rewards/chosen": -4.481690406799316,
|
| 1096 |
+
"eval_rewards/margins": 4.917874336242676,
|
| 1097 |
+
"eval_rewards/rejected": -9.399564743041992,
|
| 1098 |
+
"eval_runtime": 29.0997,
|
| 1099 |
+
"eval_samples_per_second": 27.526,
|
| 1100 |
+
"eval_steps_per_second": 3.471,
|
| 1101 |
+
"step": 600
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"epoch": 0.4924318869828456,
|
| 1105 |
+
"grad_norm": 7.072656631469727,
|
| 1106 |
+
"learning_rate": 3.222579492361179e-06,
|
| 1107 |
+
"logits/chosen": 3.2776103019714355,
|
| 1108 |
+
"logits/rejected": 3.166206121444702,
|
| 1109 |
+
"logps/chosen": -3.193716049194336,
|
| 1110 |
+
"logps/rejected": -6.185642242431641,
|
| 1111 |
+
"loss": 0.2986,
|
| 1112 |
+
"rewards/accuracies": 0.8500000238418579,
|
| 1113 |
+
"rewards/chosen": -4.790574073791504,
|
| 1114 |
+
"rewards/margins": 4.487889289855957,
|
| 1115 |
+
"rewards/rejected": -9.278463363647461,
|
| 1116 |
+
"step": 610
|
| 1117 |
+
},
|
| 1118 |
+
{
|
| 1119 |
+
"epoch": 0.5005045408678103,
|
| 1120 |
+
"grad_norm": 2.7188704013824463,
|
| 1121 |
+
"learning_rate": 3.1722995515381644e-06,
|
| 1122 |
+
"logits/chosen": 3.189532518386841,
|
| 1123 |
+
"logits/rejected": 2.9885735511779785,
|
| 1124 |
+
"logps/chosen": -3.3461785316467285,
|
| 1125 |
+
"logps/rejected": -6.720322608947754,
|
| 1126 |
+
"loss": 0.2776,
|
| 1127 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1128 |
+
"rewards/chosen": -5.019268035888672,
|
| 1129 |
+
"rewards/margins": 5.061214447021484,
|
| 1130 |
+
"rewards/rejected": -10.080483436584473,
|
| 1131 |
+
"step": 620
|
| 1132 |
+
},
|
| 1133 |
+
{
|
| 1134 |
+
"epoch": 0.508577194752775,
|
| 1135 |
+
"grad_norm": 2.6990387439727783,
|
| 1136 |
+
"learning_rate": 3.121724717912138e-06,
|
| 1137 |
+
"logits/chosen": 3.7751071453094482,
|
| 1138 |
+
"logits/rejected": 3.406925916671753,
|
| 1139 |
+
"logps/chosen": -3.3688464164733887,
|
| 1140 |
+
"logps/rejected": -6.132623672485352,
|
| 1141 |
+
"loss": 0.2688,
|
| 1142 |
+
"rewards/accuracies": 0.8374999761581421,
|
| 1143 |
+
"rewards/chosen": -5.053269386291504,
|
| 1144 |
+
"rewards/margins": 4.145665168762207,
|
| 1145 |
+
"rewards/rejected": -9.198934555053711,
|
| 1146 |
+
"step": 630
|
| 1147 |
+
},
|
| 1148 |
+
{
|
| 1149 |
+
"epoch": 0.5166498486377397,
|
| 1150 |
+
"grad_norm": 1.964301347732544,
|
| 1151 |
+
"learning_rate": 3.0708771752766397e-06,
|
| 1152 |
+
"logits/chosen": 2.6602158546447754,
|
| 1153 |
+
"logits/rejected": 2.162348508834839,
|
| 1154 |
+
"logps/chosen": -2.884718894958496,
|
| 1155 |
+
"logps/rejected": -6.869751930236816,
|
| 1156 |
+
"loss": 0.2435,
|
| 1157 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 1158 |
+
"rewards/chosen": -4.327078342437744,
|
| 1159 |
+
"rewards/margins": 5.9775495529174805,
|
| 1160 |
+
"rewards/rejected": -10.304628372192383,
|
| 1161 |
+
"step": 640
|
| 1162 |
+
},
|
| 1163 |
+
{
|
| 1164 |
+
"epoch": 0.5247225025227044,
|
| 1165 |
+
"grad_norm": 3.201704740524292,
|
| 1166 |
+
"learning_rate": 3.019779227044398e-06,
|
| 1167 |
+
"logits/chosen": 3.83849835395813,
|
| 1168 |
+
"logits/rejected": 3.5917961597442627,
|
| 1169 |
+
"logps/chosen": -3.2030792236328125,
|
| 1170 |
+
"logps/rejected": -6.711284637451172,
|
| 1171 |
+
"loss": 0.2865,
|
| 1172 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 1173 |
+
"rewards/chosen": -4.804618835449219,
|
| 1174 |
+
"rewards/margins": 5.262308120727539,
|
| 1175 |
+
"rewards/rejected": -10.066926956176758,
|
| 1176 |
+
"step": 650
|
| 1177 |
+
},
|
| 1178 |
+
{
|
| 1179 |
+
"epoch": 0.5247225025227044,
|
| 1180 |
+
"eval_logits/chosen": 2.6539251804351807,
|
| 1181 |
+
"eval_logits/rejected": 2.3508856296539307,
|
| 1182 |
+
"eval_logps/chosen": -3.0629732608795166,
|
| 1183 |
+
"eval_logps/rejected": -6.429784297943115,
|
| 1184 |
+
"eval_loss": 0.25649499893188477,
|
| 1185 |
+
"eval_rewards/accuracies": 0.9702970385551453,
|
| 1186 |
+
"eval_rewards/chosen": -4.594459533691406,
|
| 1187 |
+
"eval_rewards/margins": 5.0502166748046875,
|
| 1188 |
+
"eval_rewards/rejected": -9.644676208496094,
|
| 1189 |
+
"eval_runtime": 29.1031,
|
| 1190 |
+
"eval_samples_per_second": 27.523,
|
| 1191 |
+
"eval_steps_per_second": 3.47,
|
| 1192 |
+
"step": 650
|
| 1193 |
+
},
|
| 1194 |
+
{
|
| 1195 |
+
"epoch": 0.5327951564076691,
|
| 1196 |
+
"grad_norm": 3.517667055130005,
|
| 1197 |
+
"learning_rate": 2.9684532864643123e-06,
|
| 1198 |
+
"logits/chosen": 2.804810047149658,
|
| 1199 |
+
"logits/rejected": 2.8151614665985107,
|
| 1200 |
+
"logps/chosen": -3.4792568683624268,
|
| 1201 |
+
"logps/rejected": -7.08618688583374,
|
| 1202 |
+
"loss": 0.243,
|
| 1203 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1204 |
+
"rewards/chosen": -5.2188849449157715,
|
| 1205 |
+
"rewards/margins": 5.41039514541626,
|
| 1206 |
+
"rewards/rejected": -10.629280090332031,
|
| 1207 |
+
"step": 660
|
| 1208 |
+
},
|
| 1209 |
+
{
|
| 1210 |
+
"epoch": 0.5408678102926338,
|
| 1211 |
+
"grad_norm": 8.366011619567871,
|
| 1212 |
+
"learning_rate": 2.9169218667902562e-06,
|
| 1213 |
+
"logits/chosen": 2.1079938411712646,
|
| 1214 |
+
"logits/rejected": 2.2689383029937744,
|
| 1215 |
+
"logps/chosen": -3.3151912689208984,
|
| 1216 |
+
"logps/rejected": -7.454249382019043,
|
| 1217 |
+
"loss": 0.266,
|
| 1218 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1219 |
+
"rewards/chosen": -4.972787380218506,
|
| 1220 |
+
"rewards/margins": 6.208587169647217,
|
| 1221 |
+
"rewards/rejected": -11.181374549865723,
|
| 1222 |
+
"step": 670
|
| 1223 |
+
},
|
| 1224 |
+
{
|
| 1225 |
+
"epoch": 0.5489404641775983,
|
| 1226 |
+
"grad_norm": 4.161479473114014,
|
| 1227 |
+
"learning_rate": 2.8652075714060296e-06,
|
| 1228 |
+
"logits/chosen": 2.1858832836151123,
|
| 1229 |
+
"logits/rejected": 2.1688101291656494,
|
| 1230 |
+
"logps/chosen": -3.3753318786621094,
|
| 1231 |
+
"logps/rejected": -6.9990692138671875,
|
| 1232 |
+
"loss": 0.2356,
|
| 1233 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1234 |
+
"rewards/chosen": -5.0629987716674805,
|
| 1235 |
+
"rewards/margins": 5.435605525970459,
|
| 1236 |
+
"rewards/rejected": -10.498603820800781,
|
| 1237 |
+
"step": 680
|
| 1238 |
+
},
|
| 1239 |
+
{
|
| 1240 |
+
"epoch": 0.557013118062563,
|
| 1241 |
+
"grad_norm": 2.156621217727661,
|
| 1242 |
+
"learning_rate": 2.813333083910761e-06,
|
| 1243 |
+
"logits/chosen": 2.5004096031188965,
|
| 1244 |
+
"logits/rejected": 2.289215087890625,
|
| 1245 |
+
"logps/chosen": -3.711970567703247,
|
| 1246 |
+
"logps/rejected": -7.685518741607666,
|
| 1247 |
+
"loss": 0.2286,
|
| 1248 |
+
"rewards/accuracies": 0.9375,
|
| 1249 |
+
"rewards/chosen": -5.56795597076416,
|
| 1250 |
+
"rewards/margins": 5.960321426391602,
|
| 1251 |
+
"rewards/rejected": -11.528276443481445,
|
| 1252 |
+
"step": 690
|
| 1253 |
+
},
|
| 1254 |
+
{
|
| 1255 |
+
"epoch": 0.5650857719475277,
|
| 1256 |
+
"grad_norm": 4.547532081604004,
|
| 1257 |
+
"learning_rate": 2.761321158169134e-06,
|
| 1258 |
+
"logits/chosen": 3.112332820892334,
|
| 1259 |
+
"logits/rejected": 2.685454845428467,
|
| 1260 |
+
"logps/chosen": -3.510646343231201,
|
| 1261 |
+
"logps/rejected": -7.8125505447387695,
|
| 1262 |
+
"loss": 0.2222,
|
| 1263 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 1264 |
+
"rewards/chosen": -5.265969276428223,
|
| 1265 |
+
"rewards/margins": 6.452856540679932,
|
| 1266 |
+
"rewards/rejected": -11.718826293945312,
|
| 1267 |
+
"step": 700
|
| 1268 |
+
},
|
| 1269 |
+
{
|
| 1270 |
+
"epoch": 0.5650857719475277,
|
| 1271 |
+
"eval_logits/chosen": 2.7667236328125,
|
| 1272 |
+
"eval_logits/rejected": 2.4983408451080322,
|
| 1273 |
+
"eval_logps/chosen": -3.292154550552368,
|
| 1274 |
+
"eval_logps/rejected": -6.814427375793457,
|
| 1275 |
+
"eval_loss": 0.24892139434814453,
|
| 1276 |
+
"eval_rewards/accuracies": 0.9702970385551453,
|
| 1277 |
+
"eval_rewards/chosen": -4.938231468200684,
|
| 1278 |
+
"eval_rewards/margins": 5.2834086418151855,
|
| 1279 |
+
"eval_rewards/rejected": -10.221639633178711,
|
| 1280 |
+
"eval_runtime": 29.0862,
|
| 1281 |
+
"eval_samples_per_second": 27.539,
|
| 1282 |
+
"eval_steps_per_second": 3.472,
|
| 1283 |
+
"step": 700
|
| 1284 |
+
},
|
| 1285 |
+
{
|
| 1286 |
+
"epoch": 0.5731584258324924,
|
| 1287 |
+
"grad_norm": 3.3842532634735107,
|
| 1288 |
+
"learning_rate": 2.70919460833079e-06,
|
| 1289 |
+
"logits/chosen": 2.6863622665405273,
|
| 1290 |
+
"logits/rejected": 2.7381534576416016,
|
| 1291 |
+
"logps/chosen": -3.1926321983337402,
|
| 1292 |
+
"logps/rejected": -7.823538780212402,
|
| 1293 |
+
"loss": 0.2153,
|
| 1294 |
+
"rewards/accuracies": 0.9375,
|
| 1295 |
+
"rewards/chosen": -4.788949012756348,
|
| 1296 |
+
"rewards/margins": 6.946360111236572,
|
| 1297 |
+
"rewards/rejected": -11.735308647155762,
|
| 1298 |
+
"step": 710
|
| 1299 |
+
},
|
| 1300 |
+
{
|
| 1301 |
+
"epoch": 0.5812310797174571,
|
| 1302 |
+
"grad_norm": 2.0056378841400146,
|
| 1303 |
+
"learning_rate": 2.6569762988232838e-06,
|
| 1304 |
+
"logits/chosen": 2.938141345977783,
|
| 1305 |
+
"logits/rejected": 2.8052525520324707,
|
| 1306 |
+
"logps/chosen": -3.4339518547058105,
|
| 1307 |
+
"logps/rejected": -7.353487968444824,
|
| 1308 |
+
"loss": 0.2422,
|
| 1309 |
+
"rewards/accuracies": 0.875,
|
| 1310 |
+
"rewards/chosen": -5.150927543640137,
|
| 1311 |
+
"rewards/margins": 5.879303932189941,
|
| 1312 |
+
"rewards/rejected": -11.030232429504395,
|
| 1313 |
+
"step": 720
|
| 1314 |
+
},
|
| 1315 |
+
{
|
| 1316 |
+
"epoch": 0.5893037336024218,
|
| 1317 |
+
"grad_norm": 2.9281482696533203,
|
| 1318 |
+
"learning_rate": 2.604689134322999e-06,
|
| 1319 |
+
"logits/chosen": 3.5540454387664795,
|
| 1320 |
+
"logits/rejected": 3.6734695434570312,
|
| 1321 |
+
"logps/chosen": -3.0852925777435303,
|
| 1322 |
+
"logps/rejected": -6.869669437408447,
|
| 1323 |
+
"loss": 0.2694,
|
| 1324 |
+
"rewards/accuracies": 0.8500000238418579,
|
| 1325 |
+
"rewards/chosen": -4.627939224243164,
|
| 1326 |
+
"rewards/margins": 5.676565647125244,
|
| 1327 |
+
"rewards/rejected": -10.304505348205566,
|
| 1328 |
+
"step": 730
|
| 1329 |
+
},
|
| 1330 |
+
{
|
| 1331 |
+
"epoch": 0.5973763874873865,
|
| 1332 |
+
"grad_norm": 3.3841099739074707,
|
| 1333 |
+
"learning_rate": 2.5523560497083927e-06,
|
| 1334 |
+
"logits/chosen": 2.7171874046325684,
|
| 1335 |
+
"logits/rejected": 2.4911468029022217,
|
| 1336 |
+
"logps/chosen": -3.4943466186523438,
|
| 1337 |
+
"logps/rejected": -7.486982822418213,
|
| 1338 |
+
"loss": 0.2184,
|
| 1339 |
+
"rewards/accuracies": 0.9375,
|
| 1340 |
+
"rewards/chosen": -5.241519927978516,
|
| 1341 |
+
"rewards/margins": 5.988953113555908,
|
| 1342 |
+
"rewards/rejected": -11.230472564697266,
|
| 1343 |
+
"step": 740
|
| 1344 |
+
},
|
| 1345 |
+
{
|
| 1346 |
+
"epoch": 0.6054490413723511,
|
| 1347 |
+
"grad_norm": 2.6915197372436523,
|
| 1348 |
+
"learning_rate": 2.5e-06,
|
| 1349 |
+
"logits/chosen": 2.866816759109497,
|
| 1350 |
+
"logits/rejected": 2.84656023979187,
|
| 1351 |
+
"logps/chosen": -3.155074119567871,
|
| 1352 |
+
"logps/rejected": -8.019740104675293,
|
| 1353 |
+
"loss": 0.2364,
|
| 1354 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1355 |
+
"rewards/chosen": -4.732610702514648,
|
| 1356 |
+
"rewards/margins": 7.297000885009766,
|
| 1357 |
+
"rewards/rejected": -12.029611587524414,
|
| 1358 |
+
"step": 750
|
| 1359 |
+
},
|
| 1360 |
+
{
|
| 1361 |
+
"epoch": 0.6054490413723511,
|
| 1362 |
+
"eval_logits/chosen": 2.7122883796691895,
|
| 1363 |
+
"eval_logits/rejected": 2.4663355350494385,
|
| 1364 |
+
"eval_logps/chosen": -3.362666606903076,
|
| 1365 |
+
"eval_logps/rejected": -7.107527732849121,
|
| 1366 |
+
"eval_loss": 0.2374112904071808,
|
| 1367 |
+
"eval_rewards/accuracies": 0.9801980257034302,
|
| 1368 |
+
"eval_rewards/chosen": -5.043999671936035,
|
| 1369 |
+
"eval_rewards/margins": 5.617292881011963,
|
| 1370 |
+
"eval_rewards/rejected": -10.661293983459473,
|
| 1371 |
+
"eval_runtime": 29.0963,
|
| 1372 |
+
"eval_samples_per_second": 27.529,
|
| 1373 |
+
"eval_steps_per_second": 3.471,
|
| 1374 |
+
"step": 750
|
| 1375 |
+
},
|
| 1376 |
+
{
|
| 1377 |
+
"epoch": 0.6135216952573158,
|
| 1378 |
+
"grad_norm": 5.086276531219482,
|
| 1379 |
+
"learning_rate": 2.447643950291608e-06,
|
| 1380 |
+
"logits/chosen": 3.0712199211120605,
|
| 1381 |
+
"logits/rejected": 2.7986550331115723,
|
| 1382 |
+
"logps/chosen": -3.759767532348633,
|
| 1383 |
+
"logps/rejected": -8.20389175415039,
|
| 1384 |
+
"loss": 0.2467,
|
| 1385 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1386 |
+
"rewards/chosen": -5.639651298522949,
|
| 1387 |
+
"rewards/margins": 6.666187286376953,
|
| 1388 |
+
"rewards/rejected": -12.305837631225586,
|
| 1389 |
+
"step": 760
|
| 1390 |
+
},
|
| 1391 |
+
{
|
| 1392 |
+
"epoch": 0.6215943491422805,
|
| 1393 |
+
"grad_norm": 4.272627830505371,
|
| 1394 |
+
"learning_rate": 2.3953108656770018e-06,
|
| 1395 |
+
"logits/chosen": 2.4291179180145264,
|
| 1396 |
+
"logits/rejected": 2.247910737991333,
|
| 1397 |
+
"logps/chosen": -3.4369187355041504,
|
| 1398 |
+
"logps/rejected": -8.507123947143555,
|
| 1399 |
+
"loss": 0.2287,
|
| 1400 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 1401 |
+
"rewards/chosen": -5.155378341674805,
|
| 1402 |
+
"rewards/margins": 7.605309963226318,
|
| 1403 |
+
"rewards/rejected": -12.760688781738281,
|
| 1404 |
+
"step": 770
|
| 1405 |
+
},
|
| 1406 |
+
{
|
| 1407 |
+
"epoch": 0.6296670030272452,
|
| 1408 |
+
"grad_norm": 4.545230388641357,
|
| 1409 |
+
"learning_rate": 2.3430237011767166e-06,
|
| 1410 |
+
"logits/chosen": 3.324824810028076,
|
| 1411 |
+
"logits/rejected": 3.0623250007629395,
|
| 1412 |
+
"logps/chosen": -3.5336334705352783,
|
| 1413 |
+
"logps/rejected": -7.56264591217041,
|
| 1414 |
+
"loss": 0.259,
|
| 1415 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1416 |
+
"rewards/chosen": -5.300450325012207,
|
| 1417 |
+
"rewards/margins": 6.043517589569092,
|
| 1418 |
+
"rewards/rejected": -11.34396743774414,
|
| 1419 |
+
"step": 780
|
| 1420 |
+
},
|
| 1421 |
+
{
|
| 1422 |
+
"epoch": 0.6377396569122099,
|
| 1423 |
+
"grad_norm": 3.5774872303009033,
|
| 1424 |
+
"learning_rate": 2.290805391669212e-06,
|
| 1425 |
+
"logits/chosen": 2.70954966545105,
|
| 1426 |
+
"logits/rejected": 2.559739589691162,
|
| 1427 |
+
"logps/chosen": -3.2724106311798096,
|
| 1428 |
+
"logps/rejected": -7.964414119720459,
|
| 1429 |
+
"loss": 0.243,
|
| 1430 |
+
"rewards/accuracies": 0.9375,
|
| 1431 |
+
"rewards/chosen": -4.908616065979004,
|
| 1432 |
+
"rewards/margins": 7.0380048751831055,
|
| 1433 |
+
"rewards/rejected": -11.94662094116211,
|
| 1434 |
+
"step": 790
|
| 1435 |
+
},
|
| 1436 |
+
{
|
| 1437 |
+
"epoch": 0.6458123107971746,
|
| 1438 |
+
"grad_norm": 3.9293906688690186,
|
| 1439 |
+
"learning_rate": 2.238678841830867e-06,
|
| 1440 |
+
"logits/chosen": 2.2065200805664062,
|
| 1441 |
+
"logits/rejected": 2.5098512172698975,
|
| 1442 |
+
"logps/chosen": -3.324014186859131,
|
| 1443 |
+
"logps/rejected": -8.072491645812988,
|
| 1444 |
+
"loss": 0.2752,
|
| 1445 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 1446 |
+
"rewards/chosen": -4.986021518707275,
|
| 1447 |
+
"rewards/margins": 7.122715950012207,
|
| 1448 |
+
"rewards/rejected": -12.108736038208008,
|
| 1449 |
+
"step": 800
|
| 1450 |
+
},
|
| 1451 |
+
{
|
| 1452 |
+
"epoch": 0.6458123107971746,
|
| 1453 |
+
"eval_logits/chosen": 2.7405736446380615,
|
| 1454 |
+
"eval_logits/rejected": 2.500715732574463,
|
| 1455 |
+
"eval_logps/chosen": -3.4503068923950195,
|
| 1456 |
+
"eval_logps/rejected": -7.3543477058410645,
|
| 1457 |
+
"eval_loss": 0.22608846426010132,
|
| 1458 |
+
"eval_rewards/accuracies": 0.9801980257034302,
|
| 1459 |
+
"eval_rewards/chosen": -5.175460338592529,
|
| 1460 |
+
"eval_rewards/margins": 5.856060981750488,
|
| 1461 |
+
"eval_rewards/rejected": -11.031521797180176,
|
| 1462 |
+
"eval_runtime": 29.0926,
|
| 1463 |
+
"eval_samples_per_second": 27.533,
|
| 1464 |
+
"eval_steps_per_second": 3.472,
|
| 1465 |
+
"step": 800
|
| 1466 |
+
},
|
| 1467 |
+
{
|
| 1468 |
+
"epoch": 0.6538849646821393,
|
| 1469 |
+
"grad_norm": 3.1484081745147705,
|
| 1470 |
+
"learning_rate": 2.186666916089239e-06,
|
| 1471 |
+
"logits/chosen": 2.620433807373047,
|
| 1472 |
+
"logits/rejected": 2.554230213165283,
|
| 1473 |
+
"logps/chosen": -3.826986312866211,
|
| 1474 |
+
"logps/rejected": -8.595190048217773,
|
| 1475 |
+
"loss": 0.2267,
|
| 1476 |
+
"rewards/accuracies": 0.9375,
|
| 1477 |
+
"rewards/chosen": -5.740479469299316,
|
| 1478 |
+
"rewards/margins": 7.152307033538818,
|
| 1479 |
+
"rewards/rejected": -12.892786026000977,
|
| 1480 |
+
"step": 810
|
| 1481 |
+
},
|
| 1482 |
+
{
|
| 1483 |
+
"epoch": 0.6619576185671039,
|
| 1484 |
+
"grad_norm": 3.712259531021118,
|
| 1485 |
+
"learning_rate": 2.134792428593971e-06,
|
| 1486 |
+
"logits/chosen": 3.1829707622528076,
|
| 1487 |
+
"logits/rejected": 3.0434184074401855,
|
| 1488 |
+
"logps/chosen": -3.3385536670684814,
|
| 1489 |
+
"logps/rejected": -7.571812629699707,
|
| 1490 |
+
"loss": 0.2079,
|
| 1491 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1492 |
+
"rewards/chosen": -5.0078301429748535,
|
| 1493 |
+
"rewards/margins": 6.349888801574707,
|
| 1494 |
+
"rewards/rejected": -11.357718467712402,
|
| 1495 |
+
"step": 820
|
| 1496 |
+
},
|
| 1497 |
+
{
|
| 1498 |
+
"epoch": 0.6700302724520686,
|
| 1499 |
+
"grad_norm": 5.120253086090088,
|
| 1500 |
+
"learning_rate": 2.0830781332097446e-06,
|
| 1501 |
+
"logits/chosen": 3.631582736968994,
|
| 1502 |
+
"logits/rejected": 3.769865036010742,
|
| 1503 |
+
"logps/chosen": -3.2808995246887207,
|
| 1504 |
+
"logps/rejected": -7.530601501464844,
|
| 1505 |
+
"loss": 0.2213,
|
| 1506 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1507 |
+
"rewards/chosen": -4.92134952545166,
|
| 1508 |
+
"rewards/margins": 6.3745527267456055,
|
| 1509 |
+
"rewards/rejected": -11.29590129852295,
|
| 1510 |
+
"step": 830
|
| 1511 |
+
},
|
| 1512 |
+
{
|
| 1513 |
+
"epoch": 0.6781029263370333,
|
| 1514 |
+
"grad_norm": 3.723133087158203,
|
| 1515 |
+
"learning_rate": 2.031546713535688e-06,
|
| 1516 |
+
"logits/chosen": 2.7661056518554688,
|
| 1517 |
+
"logits/rejected": 2.7924928665161133,
|
| 1518 |
+
"logps/chosen": -3.5595524311065674,
|
| 1519 |
+
"logps/rejected": -7.967140197753906,
|
| 1520 |
+
"loss": 0.2381,
|
| 1521 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1522 |
+
"rewards/chosen": -5.339327812194824,
|
| 1523 |
+
"rewards/margins": 6.611382961273193,
|
| 1524 |
+
"rewards/rejected": -11.950711250305176,
|
| 1525 |
+
"step": 840
|
| 1526 |
+
},
|
| 1527 |
+
{
|
| 1528 |
+
"epoch": 0.686175580221998,
|
| 1529 |
+
"grad_norm": 3.4145116806030273,
|
| 1530 |
+
"learning_rate": 1.9802207729556023e-06,
|
| 1531 |
+
"logits/chosen": 2.4502460956573486,
|
| 1532 |
+
"logits/rejected": 2.364321708679199,
|
| 1533 |
+
"logps/chosen": -3.2101433277130127,
|
| 1534 |
+
"logps/rejected": -8.211054801940918,
|
| 1535 |
+
"loss": 0.2112,
|
| 1536 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 1537 |
+
"rewards/chosen": -4.81521463394165,
|
| 1538 |
+
"rewards/margins": 7.501368522644043,
|
| 1539 |
+
"rewards/rejected": -12.316584587097168,
|
| 1540 |
+
"step": 850
|
| 1541 |
+
},
|
| 1542 |
+
{
|
| 1543 |
+
"epoch": 0.686175580221998,
|
| 1544 |
+
"eval_logits/chosen": 2.7086803913116455,
|
| 1545 |
+
"eval_logits/rejected": 2.437659978866577,
|
| 1546 |
+
"eval_logps/chosen": -3.123488664627075,
|
| 1547 |
+
"eval_logps/rejected": -7.074939727783203,
|
| 1548 |
+
"eval_loss": 0.22627003490924835,
|
| 1549 |
+
"eval_rewards/accuracies": 0.9702970385551453,
|
| 1550 |
+
"eval_rewards/chosen": -4.685232639312744,
|
| 1551 |
+
"eval_rewards/margins": 5.927176475524902,
|
| 1552 |
+
"eval_rewards/rejected": -10.612408638000488,
|
| 1553 |
+
"eval_runtime": 29.0746,
|
| 1554 |
+
"eval_samples_per_second": 27.55,
|
| 1555 |
+
"eval_steps_per_second": 3.474,
|
| 1556 |
+
"step": 850
|
| 1557 |
+
},
|
| 1558 |
+
{
|
| 1559 |
+
"epoch": 0.6942482341069627,
|
| 1560 |
+
"grad_norm": 3.628746509552002,
|
| 1561 |
+
"learning_rate": 1.9291228247233607e-06,
|
| 1562 |
+
"logits/chosen": 2.9090940952301025,
|
| 1563 |
+
"logits/rejected": 2.9559710025787354,
|
| 1564 |
+
"logps/chosen": -3.447845458984375,
|
| 1565 |
+
"logps/rejected": -7.511412620544434,
|
| 1566 |
+
"loss": 0.2154,
|
| 1567 |
+
"rewards/accuracies": 0.9375,
|
| 1568 |
+
"rewards/chosen": -5.171767711639404,
|
| 1569 |
+
"rewards/margins": 6.09535026550293,
|
| 1570 |
+
"rewards/rejected": -11.267118453979492,
|
| 1571 |
+
"step": 860
|
| 1572 |
+
},
|
| 1573 |
+
{
|
| 1574 |
+
"epoch": 0.7023208879919274,
|
| 1575 |
+
"grad_norm": 3.231868267059326,
|
| 1576 |
+
"learning_rate": 1.8782752820878636e-06,
|
| 1577 |
+
"logits/chosen": 2.6312456130981445,
|
| 1578 |
+
"logits/rejected": 2.458289861679077,
|
| 1579 |
+
"logps/chosen": -3.6256003379821777,
|
| 1580 |
+
"logps/rejected": -8.079771041870117,
|
| 1581 |
+
"loss": 0.2386,
|
| 1582 |
+
"rewards/accuracies": 0.887499988079071,
|
| 1583 |
+
"rewards/chosen": -5.438401699066162,
|
| 1584 |
+
"rewards/margins": 6.681253910064697,
|
| 1585 |
+
"rewards/rejected": -12.11965560913086,
|
| 1586 |
+
"step": 870
|
| 1587 |
+
},
|
| 1588 |
+
{
|
| 1589 |
+
"epoch": 0.7103935418768921,
|
| 1590 |
+
"grad_norm": 5.360827445983887,
|
| 1591 |
+
"learning_rate": 1.827700448461836e-06,
|
| 1592 |
+
"logits/chosen": 2.383152961730957,
|
| 1593 |
+
"logits/rejected": 2.461998701095581,
|
| 1594 |
+
"logps/chosen": -3.555072069168091,
|
| 1595 |
+
"logps/rejected": -8.438495635986328,
|
| 1596 |
+
"loss": 0.2126,
|
| 1597 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1598 |
+
"rewards/chosen": -5.332608222961426,
|
| 1599 |
+
"rewards/margins": 7.32513427734375,
|
| 1600 |
+
"rewards/rejected": -12.657742500305176,
|
| 1601 |
+
"step": 880
|
| 1602 |
+
},
|
| 1603 |
+
{
|
| 1604 |
+
"epoch": 0.7184661957618567,
|
| 1605 |
+
"grad_norm": 5.990912914276123,
|
| 1606 |
+
"learning_rate": 1.7774205076388207e-06,
|
| 1607 |
+
"logits/chosen": 3.3776676654815674,
|
| 1608 |
+
"logits/rejected": 2.913318157196045,
|
| 1609 |
+
"logps/chosen": -3.9117507934570312,
|
| 1610 |
+
"logps/rejected": -8.53381633758545,
|
| 1611 |
+
"loss": 0.185,
|
| 1612 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1613 |
+
"rewards/chosen": -5.8676252365112305,
|
| 1614 |
+
"rewards/margins": 6.933099269866943,
|
| 1615 |
+
"rewards/rejected": -12.800724983215332,
|
| 1616 |
+
"step": 890
|
| 1617 |
+
},
|
| 1618 |
+
{
|
| 1619 |
+
"epoch": 0.7265388496468214,
|
| 1620 |
+
"grad_norm": 6.092798233032227,
|
| 1621 |
+
"learning_rate": 1.7274575140626318e-06,
|
| 1622 |
+
"logits/chosen": 1.917933702468872,
|
| 1623 |
+
"logits/rejected": 1.6799787282943726,
|
| 1624 |
+
"logps/chosen": -3.4144363403320312,
|
| 1625 |
+
"logps/rejected": -8.597683906555176,
|
| 1626 |
+
"loss": 0.2383,
|
| 1627 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 1628 |
+
"rewards/chosen": -5.121654987335205,
|
| 1629 |
+
"rewards/margins": 7.7748703956604,
|
| 1630 |
+
"rewards/rejected": -12.896525382995605,
|
| 1631 |
+
"step": 900
|
| 1632 |
+
},
|
| 1633 |
+
{
|
| 1634 |
+
"epoch": 0.7265388496468214,
|
| 1635 |
+
"eval_logits/chosen": 2.6657638549804688,
|
| 1636 |
+
"eval_logits/rejected": 2.4586844444274902,
|
| 1637 |
+
"eval_logps/chosen": -3.4474172592163086,
|
| 1638 |
+
"eval_logps/rejected": -7.506450176239014,
|
| 1639 |
+
"eval_loss": 0.21870465576648712,
|
| 1640 |
+
"eval_rewards/accuracies": 0.9801980257034302,
|
| 1641 |
+
"eval_rewards/chosen": -5.171125888824463,
|
| 1642 |
+
"eval_rewards/margins": 6.088548183441162,
|
| 1643 |
+
"eval_rewards/rejected": -11.259674072265625,
|
| 1644 |
+
"eval_runtime": 29.1008,
|
| 1645 |
+
"eval_samples_per_second": 27.525,
|
| 1646 |
+
"eval_steps_per_second": 3.471,
|
| 1647 |
+
"step": 900
|
| 1648 |
+
},
|
| 1649 |
+
{
|
| 1650 |
+
"epoch": 0.7346115035317861,
|
| 1651 |
+
"grad_norm": 3.623443841934204,
|
| 1652 |
+
"learning_rate": 1.677833383153542e-06,
|
| 1653 |
+
"logits/chosen": 2.915825605392456,
|
| 1654 |
+
"logits/rejected": 2.7418575286865234,
|
| 1655 |
+
"logps/chosen": -3.517927646636963,
|
| 1656 |
+
"logps/rejected": -8.27934455871582,
|
| 1657 |
+
"loss": 0.2311,
|
| 1658 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1659 |
+
"rewards/chosen": -5.276891708374023,
|
| 1660 |
+
"rewards/margins": 7.142125129699707,
|
| 1661 |
+
"rewards/rejected": -12.419015884399414,
|
| 1662 |
+
"step": 910
|
| 1663 |
+
},
|
| 1664 |
+
{
|
| 1665 |
+
"epoch": 0.7426841574167508,
|
| 1666 |
+
"grad_norm": 4.020949363708496,
|
| 1667 |
+
"learning_rate": 1.6285698816954626e-06,
|
| 1668 |
+
"logits/chosen": 2.4125845432281494,
|
| 1669 |
+
"logits/rejected": 2.6929471492767334,
|
| 1670 |
+
"logps/chosen": -3.646517276763916,
|
| 1671 |
+
"logps/rejected": -7.9424262046813965,
|
| 1672 |
+
"loss": 0.2034,
|
| 1673 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 1674 |
+
"rewards/chosen": -5.469775199890137,
|
| 1675 |
+
"rewards/margins": 6.443863868713379,
|
| 1676 |
+
"rewards/rejected": -11.913639068603516,
|
| 1677 |
+
"step": 920
|
| 1678 |
+
},
|
| 1679 |
+
{
|
| 1680 |
+
"epoch": 0.7507568113017155,
|
| 1681 |
+
"grad_norm": 4.215940475463867,
|
| 1682 |
+
"learning_rate": 1.5796886182883053e-06,
|
| 1683 |
+
"logits/chosen": 2.525031566619873,
|
| 1684 |
+
"logits/rejected": 2.478022336959839,
|
| 1685 |
+
"logps/chosen": -3.580151319503784,
|
| 1686 |
+
"logps/rejected": -8.356379508972168,
|
| 1687 |
+
"loss": 0.2301,
|
| 1688 |
+
"rewards/accuracies": 0.9375,
|
| 1689 |
+
"rewards/chosen": -5.370226860046387,
|
| 1690 |
+
"rewards/margins": 7.164343357086182,
|
| 1691 |
+
"rewards/rejected": -12.534570693969727,
|
| 1692 |
+
"step": 930
|
| 1693 |
+
},
|
| 1694 |
+
{
|
| 1695 |
+
"epoch": 0.7588294651866802,
|
| 1696 |
+
"grad_norm": 3.3429832458496094,
|
| 1697 |
+
"learning_rate": 1.5312110338697427e-06,
|
| 1698 |
+
"logits/chosen": 2.68546462059021,
|
| 1699 |
+
"logits/rejected": 2.4535439014434814,
|
| 1700 |
+
"logps/chosen": -3.716118574142456,
|
| 1701 |
+
"logps/rejected": -8.706206321716309,
|
| 1702 |
+
"loss": 0.1898,
|
| 1703 |
+
"rewards/accuracies": 0.9375,
|
| 1704 |
+
"rewards/chosen": -5.5741777420043945,
|
| 1705 |
+
"rewards/margins": 7.485130310058594,
|
| 1706 |
+
"rewards/rejected": -13.059308052062988,
|
| 1707 |
+
"step": 940
|
| 1708 |
+
},
|
| 1709 |
+
{
|
| 1710 |
+
"epoch": 0.7669021190716448,
|
| 1711 |
+
"grad_norm": 3.5254878997802734,
|
| 1712 |
+
"learning_rate": 1.4831583923105e-06,
|
| 1713 |
+
"logits/chosen": 2.924198627471924,
|
| 1714 |
+
"logits/rejected": 2.7865915298461914,
|
| 1715 |
+
"logps/chosen": -3.7330048084259033,
|
| 1716 |
+
"logps/rejected": -8.79862117767334,
|
| 1717 |
+
"loss": 0.1947,
|
| 1718 |
+
"rewards/accuracies": 0.9375,
|
| 1719 |
+
"rewards/chosen": -5.5995073318481445,
|
| 1720 |
+
"rewards/margins": 7.598424434661865,
|
| 1721 |
+
"rewards/rejected": -13.197932243347168,
|
| 1722 |
+
"step": 950
|
| 1723 |
+
},
|
| 1724 |
+
{
|
| 1725 |
+
"epoch": 0.7669021190716448,
|
| 1726 |
+
"eval_logits/chosen": 2.5911405086517334,
|
| 1727 |
+
"eval_logits/rejected": 2.4414517879486084,
|
| 1728 |
+
"eval_logps/chosen": -3.396683692932129,
|
| 1729 |
+
"eval_logps/rejected": -7.610343933105469,
|
| 1730 |
+
"eval_loss": 0.2162380963563919,
|
| 1731 |
+
"eval_rewards/accuracies": 0.9900990128517151,
|
| 1732 |
+
"eval_rewards/chosen": -5.09502649307251,
|
| 1733 |
+
"eval_rewards/margins": 6.32049036026001,
|
| 1734 |
+
"eval_rewards/rejected": -11.415514945983887,
|
| 1735 |
+
"eval_runtime": 29.0842,
|
| 1736 |
+
"eval_samples_per_second": 27.541,
|
| 1737 |
+
"eval_steps_per_second": 3.473,
|
| 1738 |
+
"step": 950
|
| 1739 |
+
},
|
| 1740 |
+
{
|
| 1741 |
+
"epoch": 0.7749747729566094,
|
| 1742 |
+
"grad_norm": 2.4495151042938232,
|
| 1743 |
+
"learning_rate": 1.4355517710873184e-06,
|
| 1744 |
+
"logits/chosen": 3.756882429122925,
|
| 1745 |
+
"logits/rejected": 3.461829423904419,
|
| 1746 |
+
"logps/chosen": -3.4112212657928467,
|
| 1747 |
+
"logps/rejected": -7.716673374176025,
|
| 1748 |
+
"loss": 0.1954,
|
| 1749 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1750 |
+
"rewards/chosen": -5.1168317794799805,
|
| 1751 |
+
"rewards/margins": 6.458177089691162,
|
| 1752 |
+
"rewards/rejected": -11.575010299682617,
|
| 1753 |
+
"step": 960
|
| 1754 |
+
},
|
| 1755 |
+
{
|
| 1756 |
+
"epoch": 0.7830474268415741,
|
| 1757 |
+
"grad_norm": 3.6960482597351074,
|
| 1758 |
+
"learning_rate": 1.388412052037682e-06,
|
| 1759 |
+
"logits/chosen": 2.683379888534546,
|
| 1760 |
+
"logits/rejected": 2.5084214210510254,
|
| 1761 |
+
"logps/chosen": -3.2802653312683105,
|
| 1762 |
+
"logps/rejected": -8.223787307739258,
|
| 1763 |
+
"loss": 0.1962,
|
| 1764 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 1765 |
+
"rewards/chosen": -4.920398712158203,
|
| 1766 |
+
"rewards/margins": 7.415283203125,
|
| 1767 |
+
"rewards/rejected": -12.33568000793457,
|
| 1768 |
+
"step": 970
|
| 1769 |
+
},
|
| 1770 |
+
{
|
| 1771 |
+
"epoch": 0.7911200807265388,
|
| 1772 |
+
"grad_norm": 1.6209157705307007,
|
| 1773 |
+
"learning_rate": 1.3417599122003464e-06,
|
| 1774 |
+
"logits/chosen": 2.2775497436523438,
|
| 1775 |
+
"logits/rejected": 2.4255974292755127,
|
| 1776 |
+
"logps/chosen": -3.296114444732666,
|
| 1777 |
+
"logps/rejected": -7.983359336853027,
|
| 1778 |
+
"loss": 0.1651,
|
| 1779 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 1780 |
+
"rewards/chosen": -4.944170951843262,
|
| 1781 |
+
"rewards/margins": 7.030867576599121,
|
| 1782 |
+
"rewards/rejected": -11.975038528442383,
|
| 1783 |
+
"step": 980
|
| 1784 |
+
},
|
| 1785 |
+
{
|
| 1786 |
+
"epoch": 0.7991927346115035,
|
| 1787 |
+
"grad_norm": 4.007768154144287,
|
| 1788 |
+
"learning_rate": 1.2956158147457116e-06,
|
| 1789 |
+
"logits/chosen": 3.2154605388641357,
|
| 1790 |
+
"logits/rejected": 2.895961046218872,
|
| 1791 |
+
"logps/chosen": -3.3481571674346924,
|
| 1792 |
+
"logps/rejected": -8.350494384765625,
|
| 1793 |
+
"loss": 0.1702,
|
| 1794 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 1795 |
+
"rewards/chosen": -5.022235870361328,
|
| 1796 |
+
"rewards/margins": 7.503505706787109,
|
| 1797 |
+
"rewards/rejected": -12.525741577148438,
|
| 1798 |
+
"step": 990
|
| 1799 |
+
},
|
| 1800 |
+
{
|
| 1801 |
+
"epoch": 0.8072653884964682,
|
| 1802 |
+
"grad_norm": 3.026970386505127,
|
| 1803 |
+
"learning_rate": 1.2500000000000007e-06,
|
| 1804 |
+
"logits/chosen": 3.6043121814727783,
|
| 1805 |
+
"logits/rejected": 3.541698932647705,
|
| 1806 |
+
"logps/chosen": -3.722698926925659,
|
| 1807 |
+
"logps/rejected": -8.218320846557617,
|
| 1808 |
+
"loss": 0.1934,
|
| 1809 |
+
"rewards/accuracies": 0.987500011920929,
|
| 1810 |
+
"rewards/chosen": -5.584048271179199,
|
| 1811 |
+
"rewards/margins": 6.743433475494385,
|
| 1812 |
+
"rewards/rejected": -12.327482223510742,
|
| 1813 |
+
"step": 1000
|
| 1814 |
+
},
|
| 1815 |
+
{
|
| 1816 |
+
"epoch": 0.8072653884964682,
|
| 1817 |
+
"eval_logits/chosen": 2.707226514816284,
|
| 1818 |
+
"eval_logits/rejected": 2.5214099884033203,
|
| 1819 |
+
"eval_logps/chosen": -3.502831220626831,
|
| 1820 |
+
"eval_logps/rejected": -7.705867767333984,
|
| 1821 |
+
"eval_loss": 0.20982278883457184,
|
| 1822 |
+
"eval_rewards/accuracies": 0.9900990128517151,
|
| 1823 |
+
"eval_rewards/chosen": -5.254247188568115,
|
| 1824 |
+
"eval_rewards/margins": 6.304553985595703,
|
| 1825 |
+
"eval_rewards/rejected": -11.558801651000977,
|
| 1826 |
+
"eval_runtime": 29.0974,
|
| 1827 |
+
"eval_samples_per_second": 27.528,
|
| 1828 |
+
"eval_steps_per_second": 3.471,
|
| 1829 |
+
"step": 1000
|
| 1830 |
+
},
|
| 1831 |
+
{
|
| 1832 |
+
"epoch": 0.8153380423814329,
|
| 1833 |
+
"grad_norm": 2.1142029762268066,
|
| 1834 |
+
"learning_rate": 1.204932476567175e-06,
|
| 1835 |
+
"logits/chosen": 3.21277117729187,
|
| 1836 |
+
"logits/rejected": 2.902198314666748,
|
| 1837 |
+
"logps/chosen": -3.9205710887908936,
|
| 1838 |
+
"logps/rejected": -8.282219886779785,
|
| 1839 |
+
"loss": 0.1911,
|
| 1840 |
+
"rewards/accuracies": 0.9375,
|
| 1841 |
+
"rewards/chosen": -5.880856513977051,
|
| 1842 |
+
"rewards/margins": 6.542473793029785,
|
| 1843 |
+
"rewards/rejected": -12.423330307006836,
|
| 1844 |
+
"step": 1010
|
| 1845 |
+
},
|
| 1846 |
+
{
|
| 1847 |
+
"epoch": 0.8234106962663976,
|
| 1848 |
+
"grad_norm": 4.392597198486328,
|
| 1849 |
+
"learning_rate": 1.160433012552508e-06,
|
| 1850 |
+
"logits/chosen": 2.785994052886963,
|
| 1851 |
+
"logits/rejected": 2.8172032833099365,
|
| 1852 |
+
"logps/chosen": -3.6294891834259033,
|
| 1853 |
+
"logps/rejected": -8.542880058288574,
|
| 1854 |
+
"loss": 0.2222,
|
| 1855 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1856 |
+
"rewards/chosen": -5.4442338943481445,
|
| 1857 |
+
"rewards/margins": 7.370086669921875,
|
| 1858 |
+
"rewards/rejected": -12.81432056427002,
|
| 1859 |
+
"step": 1020
|
| 1860 |
+
},
|
| 1861 |
+
{
|
| 1862 |
+
"epoch": 0.8314833501513622,
|
| 1863 |
+
"grad_norm": 2.633239269256592,
|
| 1864 |
+
"learning_rate": 1.11652112689164e-06,
|
| 1865 |
+
"logits/chosen": 2.663726329803467,
|
| 1866 |
+
"logits/rejected": 2.4244682788848877,
|
| 1867 |
+
"logps/chosen": -3.625394821166992,
|
| 1868 |
+
"logps/rejected": -8.94005298614502,
|
| 1869 |
+
"loss": 0.1768,
|
| 1870 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 1871 |
+
"rewards/chosen": -5.438092231750488,
|
| 1872 |
+
"rewards/margins": 7.971987247467041,
|
| 1873 |
+
"rewards/rejected": -13.410079956054688,
|
| 1874 |
+
"step": 1030
|
| 1875 |
+
},
|
| 1876 |
+
{
|
| 1877 |
+
"epoch": 0.8395560040363269,
|
| 1878 |
+
"grad_norm": 2.5595297813415527,
|
| 1879 |
+
"learning_rate": 1.073216080788921e-06,
|
| 1880 |
+
"logits/chosen": 3.432141065597534,
|
| 1881 |
+
"logits/rejected": 3.068664789199829,
|
| 1882 |
+
"logps/chosen": -3.4811911582946777,
|
| 1883 |
+
"logps/rejected": -8.416011810302734,
|
| 1884 |
+
"loss": 0.1923,
|
| 1885 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1886 |
+
"rewards/chosen": -5.221786975860596,
|
| 1887 |
+
"rewards/margins": 7.402231693267822,
|
| 1888 |
+
"rewards/rejected": -12.624017715454102,
|
| 1889 |
+
"step": 1040
|
| 1890 |
+
},
|
| 1891 |
+
{
|
| 1892 |
+
"epoch": 0.8476286579212916,
|
| 1893 |
+
"grad_norm": 5.044093132019043,
|
| 1894 |
+
"learning_rate": 1.0305368692688175e-06,
|
| 1895 |
+
"logits/chosen": 2.554561138153076,
|
| 1896 |
+
"logits/rejected": 2.273308753967285,
|
| 1897 |
+
"logps/chosen": -3.726578950881958,
|
| 1898 |
+
"logps/rejected": -8.446795463562012,
|
| 1899 |
+
"loss": 0.1525,
|
| 1900 |
+
"rewards/accuracies": 0.987500011920929,
|
| 1901 |
+
"rewards/chosen": -5.589868545532227,
|
| 1902 |
+
"rewards/margins": 7.080326080322266,
|
| 1903 |
+
"rewards/rejected": -12.670194625854492,
|
| 1904 |
+
"step": 1050
|
| 1905 |
+
},
|
| 1906 |
+
{
|
| 1907 |
+
"epoch": 0.8476286579212916,
|
| 1908 |
+
"eval_logits/chosen": 2.738980770111084,
|
| 1909 |
+
"eval_logits/rejected": 2.524784564971924,
|
| 1910 |
+
"eval_logps/chosen": -3.4412803649902344,
|
| 1911 |
+
"eval_logps/rejected": -7.715125560760498,
|
| 1912 |
+
"eval_loss": 0.20868732035160065,
|
| 1913 |
+
"eval_rewards/accuracies": 0.9801980257034302,
|
| 1914 |
+
"eval_rewards/chosen": -5.161920070648193,
|
| 1915 |
+
"eval_rewards/margins": 6.410768032073975,
|
| 1916 |
+
"eval_rewards/rejected": -11.572687149047852,
|
| 1917 |
+
"eval_runtime": 29.0996,
|
| 1918 |
+
"eval_samples_per_second": 27.526,
|
| 1919 |
+
"eval_steps_per_second": 3.471,
|
| 1920 |
+
"step": 1050
|
| 1921 |
+
},
|
| 1922 |
+
{
|
| 1923 |
+
"epoch": 0.8557013118062563,
|
| 1924 |
+
"grad_norm": 4.388390064239502,
|
| 1925 |
+
"learning_rate": 9.88502212844063e-07,
|
| 1926 |
+
"logits/chosen": 3.056361198425293,
|
| 1927 |
+
"logits/rejected": 2.8604633808135986,
|
| 1928 |
+
"logps/chosen": -3.4161269664764404,
|
| 1929 |
+
"logps/rejected": -8.637321472167969,
|
| 1930 |
+
"loss": 0.1962,
|
| 1931 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 1932 |
+
"rewards/chosen": -5.124190807342529,
|
| 1933 |
+
"rewards/margins": 7.831790924072266,
|
| 1934 |
+
"rewards/rejected": -12.955981254577637,
|
| 1935 |
+
"step": 1060
|
| 1936 |
+
},
|
| 1937 |
+
{
|
| 1938 |
+
"epoch": 0.863773965691221,
|
| 1939 |
+
"grad_norm": 5.200846195220947,
|
| 1940 |
+
"learning_rate": 9.471305493042243e-07,
|
| 1941 |
+
"logits/chosen": 2.426769971847534,
|
| 1942 |
+
"logits/rejected": 2.714139461517334,
|
| 1943 |
+
"logps/chosen": -3.3948585987091064,
|
| 1944 |
+
"logps/rejected": -8.400737762451172,
|
| 1945 |
+
"loss": 0.1953,
|
| 1946 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 1947 |
+
"rewards/chosen": -5.092287540435791,
|
| 1948 |
+
"rewards/margins": 7.508819580078125,
|
| 1949 |
+
"rewards/rejected": -12.601107597351074,
|
| 1950 |
+
"step": 1070
|
| 1951 |
+
},
|
| 1952 |
+
{
|
| 1953 |
+
"epoch": 0.8718466195761857,
|
| 1954 |
+
"grad_norm": 4.115231513977051,
|
| 1955 |
+
"learning_rate": 9.064400256282757e-07,
|
| 1956 |
+
"logits/chosen": 2.680354356765747,
|
| 1957 |
+
"logits/rejected": 2.4186928272247314,
|
| 1958 |
+
"logps/chosen": -3.406576633453369,
|
| 1959 |
+
"logps/rejected": -8.102071762084961,
|
| 1960 |
+
"loss": 0.1961,
|
| 1961 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1962 |
+
"rewards/chosen": -5.109864711761475,
|
| 1963 |
+
"rewards/margins": 7.04324197769165,
|
| 1964 |
+
"rewards/rejected": -12.153106689453125,
|
| 1965 |
+
"step": 1080
|
| 1966 |
+
},
|
| 1967 |
+
{
|
| 1968 |
+
"epoch": 0.8799192734611504,
|
| 1969 |
+
"grad_norm": 2.0617411136627197,
|
| 1970 |
+
"learning_rate": 8.664484900247363e-07,
|
| 1971 |
+
"logits/chosen": 2.8944904804229736,
|
| 1972 |
+
"logits/rejected": 2.763841152191162,
|
| 1973 |
+
"logps/chosen": -3.420555591583252,
|
| 1974 |
+
"logps/rejected": -8.535276412963867,
|
| 1975 |
+
"loss": 0.2107,
|
| 1976 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 1977 |
+
"rewards/chosen": -5.130833625793457,
|
| 1978 |
+
"rewards/margins": 7.672080993652344,
|
| 1979 |
+
"rewards/rejected": -12.802912712097168,
|
| 1980 |
+
"step": 1090
|
| 1981 |
+
},
|
| 1982 |
+
{
|
| 1983 |
+
"epoch": 0.887991927346115,
|
| 1984 |
+
"grad_norm": 4.412430763244629,
|
| 1985 |
+
"learning_rate": 8.271734841028553e-07,
|
| 1986 |
+
"logits/chosen": 2.8057053089141846,
|
| 1987 |
+
"logits/rejected": 2.880598545074463,
|
| 1988 |
+
"logps/chosen": -3.727388381958008,
|
| 1989 |
+
"logps/rejected": -8.665179252624512,
|
| 1990 |
+
"loss": 0.2103,
|
| 1991 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 1992 |
+
"rewards/chosen": -5.59108304977417,
|
| 1993 |
+
"rewards/margins": 7.406687259674072,
|
| 1994 |
+
"rewards/rejected": -12.997769355773926,
|
| 1995 |
+
"step": 1100
|
| 1996 |
+
},
|
| 1997 |
+
{
|
| 1998 |
+
"epoch": 0.887991927346115,
|
| 1999 |
+
"eval_logits/chosen": 2.784738063812256,
|
| 2000 |
+
"eval_logits/rejected": 2.529378890991211,
|
| 2001 |
+
"eval_logps/chosen": -3.4977424144744873,
|
| 2002 |
+
"eval_logps/rejected": -7.756742477416992,
|
| 2003 |
+
"eval_loss": 0.2044122815132141,
|
| 2004 |
+
"eval_rewards/accuracies": 0.9801980257034302,
|
| 2005 |
+
"eval_rewards/chosen": -5.2466139793396,
|
| 2006 |
+
"eval_rewards/margins": 6.388499736785889,
|
| 2007 |
+
"eval_rewards/rejected": -11.635113716125488,
|
| 2008 |
+
"eval_runtime": 29.0833,
|
| 2009 |
+
"eval_samples_per_second": 27.542,
|
| 2010 |
+
"eval_steps_per_second": 3.473,
|
| 2011 |
+
"step": 1100
|
| 2012 |
+
},
|
| 2013 |
+
{
|
| 2014 |
+
"epoch": 0.8960645812310797,
|
| 2015 |
+
"grad_norm": 3.327115058898926,
|
| 2016 |
+
"learning_rate": 7.886322351782782e-07,
|
| 2017 |
+
"logits/chosen": 1.8973369598388672,
|
| 2018 |
+
"logits/rejected": 1.9729961156845093,
|
| 2019 |
+
"logps/chosen": -3.3052139282226562,
|
| 2020 |
+
"logps/rejected": -8.953332901000977,
|
| 2021 |
+
"loss": 0.2092,
|
| 2022 |
+
"rewards/accuracies": 0.925000011920929,
|
| 2023 |
+
"rewards/chosen": -4.957820892333984,
|
| 2024 |
+
"rewards/margins": 8.472179412841797,
|
| 2025 |
+
"rewards/rejected": -13.430000305175781,
|
| 2026 |
+
"step": 1110
|
| 2027 |
+
},
|
| 2028 |
+
{
|
| 2029 |
+
"epoch": 0.9041372351160444,
|
| 2030 |
+
"grad_norm": 5.0973358154296875,
|
| 2031 |
+
"learning_rate": 7.508416487165862e-07,
|
| 2032 |
+
"logits/chosen": 2.4597842693328857,
|
| 2033 |
+
"logits/rejected": 2.5812220573425293,
|
| 2034 |
+
"logps/chosen": -3.381796360015869,
|
| 2035 |
+
"logps/rejected": -8.363035202026367,
|
| 2036 |
+
"loss": 0.2077,
|
| 2037 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 2038 |
+
"rewards/chosen": -5.072694301605225,
|
| 2039 |
+
"rewards/margins": 7.471858978271484,
|
| 2040 |
+
"rewards/rejected": -12.544553756713867,
|
| 2041 |
+
"step": 1120
|
| 2042 |
+
},
|
| 2043 |
+
{
|
| 2044 |
+
"epoch": 0.9122098890010091,
|
| 2045 |
+
"grad_norm": 5.166894912719727,
|
| 2046 |
+
"learning_rate": 7.138183009179922e-07,
|
| 2047 |
+
"logits/chosen": 3.5492165088653564,
|
| 2048 |
+
"logits/rejected": 3.2513375282287598,
|
| 2049 |
+
"logps/chosen": -3.6195359230041504,
|
| 2050 |
+
"logps/rejected": -8.38493824005127,
|
| 2051 |
+
"loss": 0.1904,
|
| 2052 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 2053 |
+
"rewards/chosen": -5.4293036460876465,
|
| 2054 |
+
"rewards/margins": 7.1481032371521,
|
| 2055 |
+
"rewards/rejected": -12.577406883239746,
|
| 2056 |
+
"step": 1130
|
| 2057 |
+
},
|
| 2058 |
+
{
|
| 2059 |
+
"epoch": 0.9202825428859738,
|
| 2060 |
+
"grad_norm": 3.683342695236206,
|
| 2061 |
+
"learning_rate": 6.775784314464717e-07,
|
| 2062 |
+
"logits/chosen": 2.7155990600585938,
|
| 2063 |
+
"logits/rejected": 2.7459475994110107,
|
| 2064 |
+
"logps/chosen": -4.0417985916137695,
|
| 2065 |
+
"logps/rejected": -9.647562026977539,
|
| 2066 |
+
"loss": 0.2156,
|
| 2067 |
+
"rewards/accuracies": 0.987500011920929,
|
| 2068 |
+
"rewards/chosen": -6.062697410583496,
|
| 2069 |
+
"rewards/margins": 8.40864372253418,
|
| 2070 |
+
"rewards/rejected": -14.471341133117676,
|
| 2071 |
+
"step": 1140
|
| 2072 |
+
},
|
| 2073 |
+
{
|
| 2074 |
+
"epoch": 0.9283551967709385,
|
| 2075 |
+
"grad_norm": 4.6415815353393555,
|
| 2076 |
+
"learning_rate": 6.421379363065142e-07,
|
| 2077 |
+
"logits/chosen": 2.0692670345306396,
|
| 2078 |
+
"logits/rejected": 2.390552043914795,
|
| 2079 |
+
"logps/chosen": -3.8512816429138184,
|
| 2080 |
+
"logps/rejected": -8.84770393371582,
|
| 2081 |
+
"loss": 0.2165,
|
| 2082 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 2083 |
+
"rewards/chosen": -5.776922702789307,
|
| 2084 |
+
"rewards/margins": 7.494633674621582,
|
| 2085 |
+
"rewards/rejected": -13.271554946899414,
|
| 2086 |
+
"step": 1150
|
| 2087 |
+
},
|
| 2088 |
+
{
|
| 2089 |
+
"epoch": 0.9283551967709385,
|
| 2090 |
+
"eval_logits/chosen": 2.716364860534668,
|
| 2091 |
+
"eval_logits/rejected": 2.517240047454834,
|
| 2092 |
+
"eval_logps/chosen": -3.552220344543457,
|
| 2093 |
+
"eval_logps/rejected": -7.850137710571289,
|
| 2094 |
+
"eval_loss": 0.19964782893657684,
|
| 2095 |
+
"eval_rewards/accuracies": 0.9801980257034302,
|
| 2096 |
+
"eval_rewards/chosen": -5.328330993652344,
|
| 2097 |
+
"eval_rewards/margins": 6.446875095367432,
|
| 2098 |
+
"eval_rewards/rejected": -11.775205612182617,
|
| 2099 |
+
"eval_runtime": 29.0905,
|
| 2100 |
+
"eval_samples_per_second": 27.535,
|
| 2101 |
+
"eval_steps_per_second": 3.472,
|
| 2102 |
+
"step": 1150
|
| 2103 |
+
},
|
| 2104 |
+
{
|
| 2105 |
+
"epoch": 0.9364278506559032,
|
| 2106 |
+
"grad_norm": 3.169071912765503,
|
| 2107 |
+
"learning_rate": 6.075123608706093e-07,
|
| 2108 |
+
"logits/chosen": 3.026804208755493,
|
| 2109 |
+
"logits/rejected": 2.92323637008667,
|
| 2110 |
+
"logps/chosen": -3.5037333965301514,
|
| 2111 |
+
"logps/rejected": -8.634604454040527,
|
| 2112 |
+
"loss": 0.1896,
|
| 2113 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 2114 |
+
"rewards/chosen": -5.255600929260254,
|
| 2115 |
+
"rewards/margins": 7.696307182312012,
|
| 2116 |
+
"rewards/rejected": -12.95190715789795,
|
| 2117 |
+
"step": 1160
|
| 2118 |
+
},
|
| 2119 |
+
{
|
| 2120 |
+
"epoch": 0.9445005045408678,
|
| 2121 |
+
"grad_norm": 3.537233352661133,
|
| 2122 |
+
"learning_rate": 5.737168930605272e-07,
|
| 2123 |
+
"logits/chosen": 1.8126366138458252,
|
| 2124 |
+
"logits/rejected": 2.0001204013824463,
|
| 2125 |
+
"logps/chosen": -3.8311641216278076,
|
| 2126 |
+
"logps/rejected": -8.98426342010498,
|
| 2127 |
+
"loss": 0.195,
|
| 2128 |
+
"rewards/accuracies": 0.9375,
|
| 2129 |
+
"rewards/chosen": -5.74674654006958,
|
| 2130 |
+
"rewards/margins": 7.729648590087891,
|
| 2131 |
+
"rewards/rejected": -13.476394653320312,
|
| 2132 |
+
"step": 1170
|
| 2133 |
+
},
|
| 2134 |
+
{
|
| 2135 |
+
"epoch": 0.9525731584258325,
|
| 2136 |
+
"grad_norm": 4.516322612762451,
|
| 2137 |
+
"learning_rate": 5.407663566854008e-07,
|
| 2138 |
+
"logits/chosen": 2.7332987785339355,
|
| 2139 |
+
"logits/rejected": 2.6524689197540283,
|
| 2140 |
+
"logps/chosen": -3.8542847633361816,
|
| 2141 |
+
"logps/rejected": -8.629236221313477,
|
| 2142 |
+
"loss": 0.171,
|
| 2143 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 2144 |
+
"rewards/chosen": -5.781428337097168,
|
| 2145 |
+
"rewards/margins": 7.162425994873047,
|
| 2146 |
+
"rewards/rejected": -12.943852424621582,
|
| 2147 |
+
"step": 1180
|
| 2148 |
+
},
|
| 2149 |
+
{
|
| 2150 |
+
"epoch": 0.9606458123107972,
|
| 2151 |
+
"grad_norm": 3.578998327255249,
|
| 2152 |
+
"learning_rate": 5.086752049395094e-07,
|
| 2153 |
+
"logits/chosen": 3.0999205112457275,
|
| 2154 |
+
"logits/rejected": 3.115819215774536,
|
| 2155 |
+
"logps/chosen": -3.870943784713745,
|
| 2156 |
+
"logps/rejected": -8.126202583312988,
|
| 2157 |
+
"loss": 0.2047,
|
| 2158 |
+
"rewards/accuracies": 0.9375,
|
| 2159 |
+
"rewards/chosen": -5.8064165115356445,
|
| 2160 |
+
"rewards/margins": 6.382887363433838,
|
| 2161 |
+
"rewards/rejected": -12.189303398132324,
|
| 2162 |
+
"step": 1190
|
| 2163 |
+
},
|
| 2164 |
+
{
|
| 2165 |
+
"epoch": 0.9687184661957619,
|
| 2166 |
+
"grad_norm": 4.2472405433654785,
|
| 2167 |
+
"learning_rate": 4.774575140626317e-07,
|
| 2168 |
+
"logits/chosen": 2.8590927124023438,
|
| 2169 |
+
"logits/rejected": 2.744704484939575,
|
| 2170 |
+
"logps/chosen": -4.1155290603637695,
|
| 2171 |
+
"logps/rejected": -8.967794418334961,
|
| 2172 |
+
"loss": 0.2222,
|
| 2173 |
+
"rewards/accuracies": 0.925000011920929,
|
| 2174 |
+
"rewards/chosen": -6.173293590545654,
|
| 2175 |
+
"rewards/margins": 7.2783966064453125,
|
| 2176 |
+
"rewards/rejected": -13.451690673828125,
|
| 2177 |
+
"step": 1200
|
| 2178 |
+
},
|
| 2179 |
+
{
|
| 2180 |
+
"epoch": 0.9687184661957619,
|
| 2181 |
+
"eval_logits/chosen": 2.7425291538238525,
|
| 2182 |
+
"eval_logits/rejected": 2.552403211593628,
|
| 2183 |
+
"eval_logps/chosen": -3.515270948410034,
|
| 2184 |
+
"eval_logps/rejected": -7.858344554901123,
|
| 2185 |
+
"eval_loss": 0.2018371820449829,
|
| 2186 |
+
"eval_rewards/accuracies": 0.9900990128517151,
|
| 2187 |
+
"eval_rewards/chosen": -5.27290678024292,
|
| 2188 |
+
"eval_rewards/margins": 6.514610290527344,
|
| 2189 |
+
"eval_rewards/rejected": -11.787515640258789,
|
| 2190 |
+
"eval_runtime": 29.1438,
|
| 2191 |
+
"eval_samples_per_second": 27.484,
|
| 2192 |
+
"eval_steps_per_second": 3.466,
|
| 2193 |
+
"step": 1200
|
| 2194 |
+
}
|
| 2195 |
+
],
|
| 2196 |
+
"logging_steps": 10,
|
| 2197 |
+
"max_steps": 1500,
|
| 2198 |
+
"num_input_tokens_seen": 0,
|
| 2199 |
+
"num_train_epochs": 2,
|
| 2200 |
+
"save_steps": 50,
|
| 2201 |
+
"stateful_callbacks": {
|
| 2202 |
+
"TrainerControl": {
|
| 2203 |
+
"args": {
|
| 2204 |
+
"should_epoch_stop": false,
|
| 2205 |
+
"should_evaluate": false,
|
| 2206 |
+
"should_log": false,
|
| 2207 |
+
"should_save": true,
|
| 2208 |
+
"should_training_stop": false
|
| 2209 |
+
},
|
| 2210 |
+
"attributes": {}
|
| 2211 |
+
}
|
| 2212 |
+
},
|
| 2213 |
+
"total_flos": 2.7938144021412577e+18,
|
| 2214 |
+
"train_batch_size": 1,
|
| 2215 |
+
"trial_name": null,
|
| 2216 |
+
"trial_params": null
|
| 2217 |
+
}
|
checkpoint-1200/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:448b13b23e63cb2ff5ca633312a3af2a8ddf68feacd12942d98b8afe864faf6a
|
| 3 |
+
size 7224
|
checkpoint-1200/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)
|