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 |
+
"qkv_proj",
|
| 27 |
+
"gate_up_proj",
|
| 28 |
+
"down_proj",
|
| 29 |
+
"o_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 sha256:13eac32a7e5d4a007ae1f6859a2d7606e91e1f0a1d2056d3faeb17e3f9dbd074
|
| 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:bddccb77b9758cb4bf1d2d18a966ddd1f99b4a5cd24755bbfd986c37bb7f878b
|
| 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 sha256:16847c5352b3c0f183cb340f5d2966abb3030797dd33d5cb208aa0a6ad228de4
|
| 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:f5ce0ebd68ece191eac4b67d470a64fa8b9a1eed97104cd62df01d8bb1ca2c56
|
| 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 sha256:8957a6a1c81fb4f89857f6977ebaa7f3bb2c84022fd790edfebf482b9ae56318
|
| 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 sha256:9e072a5f227102400e2abe45fa10775ca66cb0f274a1da693b4bde04bec97f44
|
| 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 sha256:03454f67fca2fe018ab5a421cf7682c5e06c7e0ad308064e3249db05cfcbb621
|
| 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 sha256:5f368bffa5024bd46ab904a6ffebdf1a38ec10cb2986d8955a927f8df741961f
|
| 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 sha256:9f8f5402f97e74dd1b797912c12fbe917c7a85ebbb8ebd1b7832b631865754ac
|
| 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 sha256:01abc7e7914ddf24a79aaa4d97304082a00b58a95f25cf1f5820c3200c7cb42e
|
| 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 sha256:479e628a41471621ed71e5888d4bc3eecd148e78c535bc6314d7d47b1c46d139
|
| 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 sha256:977d4030d8e6485cb5e6f8932d356d077b621fc5e3a3a17c1cd7300e3c3aa378
|
| 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:f6cee236c0b6ea406dfd78dc5d74c5d225be0fa30781914f3736a59b5bf50da8
|
| 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 sha256:151976a0e4efc5762247214e2e9304a5c082e878962f2b322e00eb30ef03cab3
|
| 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:ea64ac8295f29f6c548b206fafc96dea4fe6bf7d44797d6d0470afa88b072460
|
| 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:f61787141f0f1134f74a65e87d4d3ea74e81f6b9d697755ee30ef0b88b18ddba
|
| 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 sha256:787193a355e21da1f4d95b759a157a105a86b3062286f0d9df1eb7fd2c5cd54c
|
| 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:c8b1b9e4e9a20f686ee92a48f0a8089eb687a9e2a7b8be07f71fdf84e2fcf1a4
|
| 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": 1.279317697228145,
|
| 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": 0.010660980810234541,
|
| 13 |
+
"grad_norm": 0.051327500492334366,
|
| 14 |
+
"learning_rate": 4.999451708687114e-06,
|
| 15 |
+
"logits/chosen": 14.755006790161133,
|
| 16 |
+
"logits/rejected": 14.735244750976562,
|
| 17 |
+
"logps/chosen": -0.29377540946006775,
|
| 18 |
+
"logps/rejected": -0.30969956517219543,
|
| 19 |
+
"loss": 0.952,
|
| 20 |
+
"rewards/accuracies": 0.4375,
|
| 21 |
+
"rewards/chosen": -0.44066309928894043,
|
| 22 |
+
"rewards/margins": 0.023886267095804214,
|
| 23 |
+
"rewards/rejected": -0.46454939246177673,
|
| 24 |
+
"step": 10
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.021321961620469083,
|
| 28 |
+
"grad_norm": 0.04346882924437523,
|
| 29 |
+
"learning_rate": 4.997807075247147e-06,
|
| 30 |
+
"logits/chosen": 14.513801574707031,
|
| 31 |
+
"logits/rejected": 14.946454048156738,
|
| 32 |
+
"logps/chosen": -0.27995699644088745,
|
| 33 |
+
"logps/rejected": -0.30138006806373596,
|
| 34 |
+
"loss": 0.9726,
|
| 35 |
+
"rewards/accuracies": 0.4124999940395355,
|
| 36 |
+
"rewards/chosen": -0.4199354648590088,
|
| 37 |
+
"rewards/margins": 0.03213457390666008,
|
| 38 |
+
"rewards/rejected": -0.45207005739212036,
|
| 39 |
+
"step": 20
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.031982942430703626,
|
| 43 |
+
"grad_norm": 0.05228634551167488,
|
| 44 |
+
"learning_rate": 4.9950668210706795e-06,
|
| 45 |
+
"logits/chosen": 14.266324043273926,
|
| 46 |
+
"logits/rejected": 14.423965454101562,
|
| 47 |
+
"logps/chosen": -0.2919609546661377,
|
| 48 |
+
"logps/rejected": -0.32358455657958984,
|
| 49 |
+
"loss": 0.9622,
|
| 50 |
+
"rewards/accuracies": 0.5,
|
| 51 |
+
"rewards/chosen": -0.43794146180152893,
|
| 52 |
+
"rewards/margins": 0.047435395419597626,
|
| 53 |
+
"rewards/rejected": -0.48537683486938477,
|
| 54 |
+
"step": 30
|
| 55 |
+
},
|
| 56 |
+
{
|
| 57 |
+
"epoch": 0.042643923240938165,
|
| 58 |
+
"grad_norm": 0.05487598106265068,
|
| 59 |
+
"learning_rate": 4.9912321481237616e-06,
|
| 60 |
+
"logits/chosen": 14.965211868286133,
|
| 61 |
+
"logits/rejected": 15.058088302612305,
|
| 62 |
+
"logps/chosen": -0.277716726064682,
|
| 63 |
+
"logps/rejected": -0.3055034577846527,
|
| 64 |
+
"loss": 0.9403,
|
| 65 |
+
"rewards/accuracies": 0.4000000059604645,
|
| 66 |
+
"rewards/chosen": -0.4165751039981842,
|
| 67 |
+
"rewards/margins": 0.04168009012937546,
|
| 68 |
+
"rewards/rejected": -0.4582551419734955,
|
| 69 |
+
"step": 40
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 0.053304904051172705,
|
| 73 |
+
"grad_norm": 0.057255037128925323,
|
| 74 |
+
"learning_rate": 4.986304738420684e-06,
|
| 75 |
+
"logits/chosen": 14.539288520812988,
|
| 76 |
+
"logits/rejected": 15.174041748046875,
|
| 77 |
+
"logps/chosen": -0.26362231373786926,
|
| 78 |
+
"logps/rejected": -0.3325727581977844,
|
| 79 |
+
"loss": 0.9588,
|
| 80 |
+
"rewards/accuracies": 0.4749999940395355,
|
| 81 |
+
"rewards/chosen": -0.3954334557056427,
|
| 82 |
+
"rewards/margins": 0.10342560708522797,
|
| 83 |
+
"rewards/rejected": -0.49885907769203186,
|
| 84 |
+
"step": 50
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 0.053304904051172705,
|
| 88 |
+
"eval_logits/chosen": 14.618952751159668,
|
| 89 |
+
"eval_logits/rejected": 15.176809310913086,
|
| 90 |
+
"eval_logps/chosen": -0.2685677409172058,
|
| 91 |
+
"eval_logps/rejected": -0.3283654451370239,
|
| 92 |
+
"eval_loss": 0.9551004767417908,
|
| 93 |
+
"eval_rewards/accuracies": 0.5131579041481018,
|
| 94 |
+
"eval_rewards/chosen": -0.4028516113758087,
|
| 95 |
+
"eval_rewards/margins": 0.08969658613204956,
|
| 96 |
+
"eval_rewards/rejected": -0.4925481975078583,
|
| 97 |
+
"eval_runtime": 21.4453,
|
| 98 |
+
"eval_samples_per_second": 28.305,
|
| 99 |
+
"eval_steps_per_second": 3.544,
|
| 100 |
+
"step": 50
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.06396588486140725,
|
| 104 |
+
"grad_norm": 0.05227242782711983,
|
| 105 |
+
"learning_rate": 4.980286753286196e-06,
|
| 106 |
+
"logits/chosen": 14.787714004516602,
|
| 107 |
+
"logits/rejected": 15.379422187805176,
|
| 108 |
+
"logps/chosen": -0.3143109679222107,
|
| 109 |
+
"logps/rejected": -0.3425135612487793,
|
| 110 |
+
"loss": 0.9636,
|
| 111 |
+
"rewards/accuracies": 0.4625000059604645,
|
| 112 |
+
"rewards/chosen": -0.4714665412902832,
|
| 113 |
+
"rewards/margins": 0.042303841561079025,
|
| 114 |
+
"rewards/rejected": -0.513770341873169,
|
| 115 |
+
"step": 60
|
| 116 |
+
},
|
| 117 |
+
{
|
| 118 |
+
"epoch": 0.07462686567164178,
|
| 119 |
+
"grad_norm": 0.0658508762717247,
|
| 120 |
+
"learning_rate": 4.973180832407471e-06,
|
| 121 |
+
"logits/chosen": 15.149365425109863,
|
| 122 |
+
"logits/rejected": 15.115835189819336,
|
| 123 |
+
"logps/chosen": -0.31501108407974243,
|
| 124 |
+
"logps/rejected": -0.2854115962982178,
|
| 125 |
+
"loss": 0.9677,
|
| 126 |
+
"rewards/accuracies": 0.36250001192092896,
|
| 127 |
+
"rewards/chosen": -0.47251659631729126,
|
| 128 |
+
"rewards/margins": -0.04439922422170639,
|
| 129 |
+
"rewards/rejected": -0.4281173646450043,
|
| 130 |
+
"step": 70
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 0.08528784648187633,
|
| 134 |
+
"grad_norm": 0.06567618995904922,
|
| 135 |
+
"learning_rate": 4.964990092676263e-06,
|
| 136 |
+
"logits/chosen": 15.393908500671387,
|
| 137 |
+
"logits/rejected": 15.454248428344727,
|
| 138 |
+
"logps/chosen": -0.31166282296180725,
|
| 139 |
+
"logps/rejected": -0.3178747594356537,
|
| 140 |
+
"loss": 0.9609,
|
| 141 |
+
"rewards/accuracies": 0.42500001192092896,
|
| 142 |
+
"rewards/chosen": -0.4674941897392273,
|
| 143 |
+
"rewards/margins": 0.009317949414253235,
|
| 144 |
+
"rewards/rejected": -0.47681212425231934,
|
| 145 |
+
"step": 80
|
| 146 |
+
},
|
| 147 |
+
{
|
| 148 |
+
"epoch": 0.09594882729211088,
|
| 149 |
+
"grad_norm": 0.07566913962364197,
|
| 150 |
+
"learning_rate": 4.9557181268217225e-06,
|
| 151 |
+
"logits/chosen": 15.229632377624512,
|
| 152 |
+
"logits/rejected": 15.477168083190918,
|
| 153 |
+
"logps/chosen": -0.3294064998626709,
|
| 154 |
+
"logps/rejected": -0.3528878390789032,
|
| 155 |
+
"loss": 0.9587,
|
| 156 |
+
"rewards/accuracies": 0.4749999940395355,
|
| 157 |
+
"rewards/chosen": -0.49410971999168396,
|
| 158 |
+
"rewards/margins": 0.03522203490138054,
|
| 159 |
+
"rewards/rejected": -0.5293318033218384,
|
| 160 |
+
"step": 90
|
| 161 |
+
},
|
| 162 |
+
{
|
| 163 |
+
"epoch": 0.10660980810234541,
|
| 164 |
+
"grad_norm": 0.09082464128732681,
|
| 165 |
+
"learning_rate": 4.9453690018345144e-06,
|
| 166 |
+
"logits/chosen": 14.481330871582031,
|
| 167 |
+
"logits/rejected": 15.092982292175293,
|
| 168 |
+
"logps/chosen": -0.2656436562538147,
|
| 169 |
+
"logps/rejected": -0.33982905745506287,
|
| 170 |
+
"loss": 0.9548,
|
| 171 |
+
"rewards/accuracies": 0.5625,
|
| 172 |
+
"rewards/chosen": -0.39846545457839966,
|
| 173 |
+
"rewards/margins": 0.11127817630767822,
|
| 174 |
+
"rewards/rejected": -0.5097435712814331,
|
| 175 |
+
"step": 100
|
| 176 |
+
},
|
| 177 |
+
{
|
| 178 |
+
"epoch": 0.10660980810234541,
|
| 179 |
+
"eval_logits/chosen": 14.7100830078125,
|
| 180 |
+
"eval_logits/rejected": 15.274725914001465,
|
| 181 |
+
"eval_logps/chosen": -0.26462864875793457,
|
| 182 |
+
"eval_logps/rejected": -0.331702321767807,
|
| 183 |
+
"eval_loss": 0.947841465473175,
|
| 184 |
+
"eval_rewards/accuracies": 0.5394737124443054,
|
| 185 |
+
"eval_rewards/chosen": -0.39694297313690186,
|
| 186 |
+
"eval_rewards/margins": 0.10061051696538925,
|
| 187 |
+
"eval_rewards/rejected": -0.4975534677505493,
|
| 188 |
+
"eval_runtime": 21.4421,
|
| 189 |
+
"eval_samples_per_second": 28.309,
|
| 190 |
+
"eval_steps_per_second": 3.544,
|
| 191 |
+
"step": 100
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.11727078891257996,
|
| 195 |
+
"grad_norm": 0.20198923349380493,
|
| 196 |
+
"learning_rate": 4.933947257182901e-06,
|
| 197 |
+
"logits/chosen": 14.932653427124023,
|
| 198 |
+
"logits/rejected": 15.476409912109375,
|
| 199 |
+
"logps/chosen": -0.27830976247787476,
|
| 200 |
+
"logps/rejected": -0.34150317311286926,
|
| 201 |
+
"loss": 0.9487,
|
| 202 |
+
"rewards/accuracies": 0.550000011920929,
|
| 203 |
+
"rewards/chosen": -0.41746464371681213,
|
| 204 |
+
"rewards/margins": 0.09479012340307236,
|
| 205 |
+
"rewards/rejected": -0.5122548341751099,
|
| 206 |
+
"step": 110
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.1279317697228145,
|
| 210 |
+
"grad_norm": 0.31938356161117554,
|
| 211 |
+
"learning_rate": 4.921457902821578e-06,
|
| 212 |
+
"logits/chosen": 15.280967712402344,
|
| 213 |
+
"logits/rejected": 15.5416259765625,
|
| 214 |
+
"logps/chosen": -0.2816022038459778,
|
| 215 |
+
"logps/rejected": -0.3262938857078552,
|
| 216 |
+
"loss": 0.9483,
|
| 217 |
+
"rewards/accuracies": 0.4749999940395355,
|
| 218 |
+
"rewards/chosen": -0.4224032461643219,
|
| 219 |
+
"rewards/margins": 0.06703753769397736,
|
| 220 |
+
"rewards/rejected": -0.48944082856178284,
|
| 221 |
+
"step": 120
|
| 222 |
+
},
|
| 223 |
+
{
|
| 224 |
+
"epoch": 0.13859275053304904,
|
| 225 |
+
"grad_norm": 0.12567812204360962,
|
| 226 |
+
"learning_rate": 4.907906416994146e-06,
|
| 227 |
+
"logits/chosen": 14.967382431030273,
|
| 228 |
+
"logits/rejected": 15.351877212524414,
|
| 229 |
+
"logps/chosen": -0.3148510456085205,
|
| 230 |
+
"logps/rejected": -0.3488944172859192,
|
| 231 |
+
"loss": 0.957,
|
| 232 |
+
"rewards/accuracies": 0.48750001192092896,
|
| 233 |
+
"rewards/chosen": -0.47227659821510315,
|
| 234 |
+
"rewards/margins": 0.05106503888964653,
|
| 235 |
+
"rewards/rejected": -0.5233416557312012,
|
| 236 |
+
"step": 130
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 0.14925373134328357,
|
| 240 |
+
"grad_norm": 0.09151162207126617,
|
| 241 |
+
"learning_rate": 4.893298743830168e-06,
|
| 242 |
+
"logits/chosen": 14.900466918945312,
|
| 243 |
+
"logits/rejected": 15.075350761413574,
|
| 244 |
+
"logps/chosen": -0.2766302227973938,
|
| 245 |
+
"logps/rejected": -0.312236487865448,
|
| 246 |
+
"loss": 0.9373,
|
| 247 |
+
"rewards/accuracies": 0.42500001192092896,
|
| 248 |
+
"rewards/chosen": -0.4149452745914459,
|
| 249 |
+
"rewards/margins": 0.05340944975614548,
|
| 250 |
+
"rewards/rejected": -0.4683547616004944,
|
| 251 |
+
"step": 140
|
| 252 |
+
},
|
| 253 |
+
{
|
| 254 |
+
"epoch": 0.15991471215351813,
|
| 255 |
+
"grad_norm": 0.1259378045797348,
|
| 256 |
+
"learning_rate": 4.8776412907378845e-06,
|
| 257 |
+
"logits/chosen": 14.528109550476074,
|
| 258 |
+
"logits/rejected": 14.861102104187012,
|
| 259 |
+
"logps/chosen": -0.2683579921722412,
|
| 260 |
+
"logps/rejected": -0.33838269114494324,
|
| 261 |
+
"loss": 0.9388,
|
| 262 |
+
"rewards/accuracies": 0.5375000238418579,
|
| 263 |
+
"rewards/chosen": -0.40253695845603943,
|
| 264 |
+
"rewards/margins": 0.10503707826137543,
|
| 265 |
+
"rewards/rejected": -0.5075740218162537,
|
| 266 |
+
"step": 150
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 0.15991471215351813,
|
| 270 |
+
"eval_logits/chosen": 14.12246036529541,
|
| 271 |
+
"eval_logits/rejected": 14.733266830444336,
|
| 272 |
+
"eval_logps/chosen": -0.2611957788467407,
|
| 273 |
+
"eval_logps/rejected": -0.3492279350757599,
|
| 274 |
+
"eval_loss": 0.9302574396133423,
|
| 275 |
+
"eval_rewards/accuracies": 0.5657894611358643,
|
| 276 |
+
"eval_rewards/chosen": -0.3917936384677887,
|
| 277 |
+
"eval_rewards/margins": 0.13204820454120636,
|
| 278 |
+
"eval_rewards/rejected": -0.5238418579101562,
|
| 279 |
+
"eval_runtime": 21.4406,
|
| 280 |
+
"eval_samples_per_second": 28.311,
|
| 281 |
+
"eval_steps_per_second": 3.545,
|
| 282 |
+
"step": 150
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 0.17057569296375266,
|
| 286 |
+
"grad_norm": 0.11400051414966583,
|
| 287 |
+
"learning_rate": 4.860940925593703e-06,
|
| 288 |
+
"logits/chosen": 14.4571533203125,
|
| 289 |
+
"logits/rejected": 14.769159317016602,
|
| 290 |
+
"logps/chosen": -0.31032469868659973,
|
| 291 |
+
"logps/rejected": -0.34650668501853943,
|
| 292 |
+
"loss": 0.9396,
|
| 293 |
+
"rewards/accuracies": 0.512499988079071,
|
| 294 |
+
"rewards/chosen": -0.4654870927333832,
|
| 295 |
+
"rewards/margins": 0.05427298694849014,
|
| 296 |
+
"rewards/rejected": -0.519760012626648,
|
| 297 |
+
"step": 160
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 0.1812366737739872,
|
| 301 |
+
"grad_norm": 0.1102401539683342,
|
| 302 |
+
"learning_rate": 4.84320497372973e-06,
|
| 303 |
+
"logits/chosen": 13.959765434265137,
|
| 304 |
+
"logits/rejected": 14.27458381652832,
|
| 305 |
+
"logps/chosen": -0.2744378447532654,
|
| 306 |
+
"logps/rejected": -0.35702812671661377,
|
| 307 |
+
"loss": 0.9222,
|
| 308 |
+
"rewards/accuracies": 0.5375000238418579,
|
| 309 |
+
"rewards/chosen": -0.41165676712989807,
|
| 310 |
+
"rewards/margins": 0.12388546764850616,
|
| 311 |
+
"rewards/rejected": -0.5355421900749207,
|
| 312 |
+
"step": 170
|
| 313 |
+
},
|
| 314 |
+
{
|
| 315 |
+
"epoch": 0.19189765458422176,
|
| 316 |
+
"grad_norm": 0.14721031486988068,
|
| 317 |
+
"learning_rate": 4.824441214720629e-06,
|
| 318 |
+
"logits/chosen": 13.54602336883545,
|
| 319 |
+
"logits/rejected": 14.076690673828125,
|
| 320 |
+
"logps/chosen": -0.2713850140571594,
|
| 321 |
+
"logps/rejected": -0.40618976950645447,
|
| 322 |
+
"loss": 0.9052,
|
| 323 |
+
"rewards/accuracies": 0.675000011920929,
|
| 324 |
+
"rewards/chosen": -0.40707746148109436,
|
| 325 |
+
"rewards/margins": 0.20220720767974854,
|
| 326 |
+
"rewards/rejected": -0.6092846989631653,
|
| 327 |
+
"step": 180
|
| 328 |
+
},
|
| 329 |
+
{
|
| 330 |
+
"epoch": 0.2025586353944563,
|
| 331 |
+
"grad_norm": 0.1756824553012848,
|
| 332 |
+
"learning_rate": 4.804657878971252e-06,
|
| 333 |
+
"logits/chosen": 12.6314697265625,
|
| 334 |
+
"logits/rejected": 13.246849060058594,
|
| 335 |
+
"logps/chosen": -0.27216213941574097,
|
| 336 |
+
"logps/rejected": -0.4351380467414856,
|
| 337 |
+
"loss": 0.8996,
|
| 338 |
+
"rewards/accuracies": 0.6875,
|
| 339 |
+
"rewards/chosen": -0.40824323892593384,
|
| 340 |
+
"rewards/margins": 0.24446387588977814,
|
| 341 |
+
"rewards/rejected": -0.652707040309906,
|
| 342 |
+
"step": 190
|
| 343 |
+
},
|
| 344 |
+
{
|
| 345 |
+
"epoch": 0.21321961620469082,
|
| 346 |
+
"grad_norm": 0.15476027131080627,
|
| 347 |
+
"learning_rate": 4.783863644106502e-06,
|
| 348 |
+
"logits/chosen": 12.40199089050293,
|
| 349 |
+
"logits/rejected": 12.966108322143555,
|
| 350 |
+
"logps/chosen": -0.303610622882843,
|
| 351 |
+
"logps/rejected": -0.423031747341156,
|
| 352 |
+
"loss": 0.9015,
|
| 353 |
+
"rewards/accuracies": 0.5375000238418579,
|
| 354 |
+
"rewards/chosen": -0.4554159641265869,
|
| 355 |
+
"rewards/margins": 0.17913168668746948,
|
| 356 |
+
"rewards/rejected": -0.6345476508140564,
|
| 357 |
+
"step": 200
|
| 358 |
+
},
|
| 359 |
+
{
|
| 360 |
+
"epoch": 0.21321961620469082,
|
| 361 |
+
"eval_logits/chosen": 11.887229919433594,
|
| 362 |
+
"eval_logits/rejected": 12.5900239944458,
|
| 363 |
+
"eval_logps/chosen": -0.269090861082077,
|
| 364 |
+
"eval_logps/rejected": -0.42408913373947144,
|
| 365 |
+
"eval_loss": 0.8796805143356323,
|
| 366 |
+
"eval_rewards/accuracies": 0.6447368264198303,
|
| 367 |
+
"eval_rewards/chosen": -0.40363630652427673,
|
| 368 |
+
"eval_rewards/margins": 0.23249731957912445,
|
| 369 |
+
"eval_rewards/rejected": -0.6361336708068848,
|
| 370 |
+
"eval_runtime": 21.4455,
|
| 371 |
+
"eval_samples_per_second": 28.304,
|
| 372 |
+
"eval_steps_per_second": 3.544,
|
| 373 |
+
"step": 200
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.22388059701492538,
|
| 377 |
+
"grad_norm": 0.18212148547172546,
|
| 378 |
+
"learning_rate": 4.762067631165049e-06,
|
| 379 |
+
"logits/chosen": 12.375594139099121,
|
| 380 |
+
"logits/rejected": 12.701678276062012,
|
| 381 |
+
"logps/chosen": -0.3136894702911377,
|
| 382 |
+
"logps/rejected": -0.3944609761238098,
|
| 383 |
+
"loss": 0.8898,
|
| 384 |
+
"rewards/accuracies": 0.4625000059604645,
|
| 385 |
+
"rewards/chosen": -0.47053417563438416,
|
| 386 |
+
"rewards/margins": 0.12115727365016937,
|
| 387 |
+
"rewards/rejected": -0.5916914939880371,
|
| 388 |
+
"step": 210
|
| 389 |
+
},
|
| 390 |
+
{
|
| 391 |
+
"epoch": 0.2345415778251599,
|
| 392 |
+
"grad_norm": 0.5440058708190918,
|
| 393 |
+
"learning_rate": 4.7392794005985324e-06,
|
| 394 |
+
"logits/chosen": 11.23914909362793,
|
| 395 |
+
"logits/rejected": 11.926396369934082,
|
| 396 |
+
"logps/chosen": -0.3077571392059326,
|
| 397 |
+
"logps/rejected": -0.43772149085998535,
|
| 398 |
+
"loss": 0.8806,
|
| 399 |
+
"rewards/accuracies": 0.5375000238418579,
|
| 400 |
+
"rewards/chosen": -0.4616357684135437,
|
| 401 |
+
"rewards/margins": 0.19494646787643433,
|
| 402 |
+
"rewards/rejected": -0.656582236289978,
|
| 403 |
+
"step": 220
|
| 404 |
+
},
|
| 405 |
+
{
|
| 406 |
+
"epoch": 0.24520255863539445,
|
| 407 |
+
"grad_norm": 0.5628307461738586,
|
| 408 |
+
"learning_rate": 4.715508948078037e-06,
|
| 409 |
+
"logits/chosen": 11.177714347839355,
|
| 410 |
+
"logits/rejected": 11.534266471862793,
|
| 411 |
+
"logps/chosen": -0.31991320848464966,
|
| 412 |
+
"logps/rejected": -0.4394511282444,
|
| 413 |
+
"loss": 0.8778,
|
| 414 |
+
"rewards/accuracies": 0.550000011920929,
|
| 415 |
+
"rewards/chosen": -0.4798697829246521,
|
| 416 |
+
"rewards/margins": 0.17930689454078674,
|
| 417 |
+
"rewards/rejected": -0.6591767072677612,
|
| 418 |
+
"step": 230
|
| 419 |
+
},
|
| 420 |
+
{
|
| 421 |
+
"epoch": 0.255863539445629,
|
| 422 |
+
"grad_norm": 0.40485626459121704,
|
| 423 |
+
"learning_rate": 4.690766700109659e-06,
|
| 424 |
+
"logits/chosen": 10.132668495178223,
|
| 425 |
+
"logits/rejected": 10.29063606262207,
|
| 426 |
+
"logps/chosen": -0.3195653557777405,
|
| 427 |
+
"logps/rejected": -0.47949132323265076,
|
| 428 |
+
"loss": 0.8551,
|
| 429 |
+
"rewards/accuracies": 0.5874999761581421,
|
| 430 |
+
"rewards/chosen": -0.47934800386428833,
|
| 431 |
+
"rewards/margins": 0.23988890647888184,
|
| 432 |
+
"rewards/rejected": -0.7192369699478149,
|
| 433 |
+
"step": 240
|
| 434 |
+
},
|
| 435 |
+
{
|
| 436 |
+
"epoch": 0.26652452025586354,
|
| 437 |
+
"grad_norm": 0.6199322938919067,
|
| 438 |
+
"learning_rate": 4.665063509461098e-06,
|
| 439 |
+
"logits/chosen": 8.781888008117676,
|
| 440 |
+
"logits/rejected": 9.237382888793945,
|
| 441 |
+
"logps/chosen": -0.3370448052883148,
|
| 442 |
+
"logps/rejected": -0.610824465751648,
|
| 443 |
+
"loss": 0.8416,
|
| 444 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 445 |
+
"rewards/chosen": -0.505567193031311,
|
| 446 |
+
"rewards/margins": 0.4106695055961609,
|
| 447 |
+
"rewards/rejected": -0.9162367582321167,
|
| 448 |
+
"step": 250
|
| 449 |
+
},
|
| 450 |
+
{
|
| 451 |
+
"epoch": 0.26652452025586354,
|
| 452 |
+
"eval_logits/chosen": 8.437722206115723,
|
| 453 |
+
"eval_logits/rejected": 8.843962669372559,
|
| 454 |
+
"eval_logps/chosen": -0.3058585226535797,
|
| 455 |
+
"eval_logps/rejected": -0.582990825176239,
|
| 456 |
+
"eval_loss": 0.8036603331565857,
|
| 457 |
+
"eval_rewards/accuracies": 0.6447368264198303,
|
| 458 |
+
"eval_rewards/chosen": -0.4587877094745636,
|
| 459 |
+
"eval_rewards/margins": 0.4156985878944397,
|
| 460 |
+
"eval_rewards/rejected": -0.8744862079620361,
|
| 461 |
+
"eval_runtime": 21.4423,
|
| 462 |
+
"eval_samples_per_second": 28.308,
|
| 463 |
+
"eval_steps_per_second": 3.544,
|
| 464 |
+
"step": 250
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.2771855010660981,
|
| 468 |
+
"grad_norm": 0.3213505744934082,
|
| 469 |
+
"learning_rate": 4.638410650401267e-06,
|
| 470 |
+
"logits/chosen": 7.914826393127441,
|
| 471 |
+
"logits/rejected": 8.010818481445312,
|
| 472 |
+
"logps/chosen": -0.3556877374649048,
|
| 473 |
+
"logps/rejected": -0.7540119886398315,
|
| 474 |
+
"loss": 0.7811,
|
| 475 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 476 |
+
"rewards/chosen": -0.5335315465927124,
|
| 477 |
+
"rewards/margins": 0.5974863171577454,
|
| 478 |
+
"rewards/rejected": -1.1310179233551025,
|
| 479 |
+
"step": 260
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 0.2878464818763326,
|
| 483 |
+
"grad_norm": 1.0119378566741943,
|
| 484 |
+
"learning_rate": 4.610819813755038e-06,
|
| 485 |
+
"logits/chosen": 7.584845542907715,
|
| 486 |
+
"logits/rejected": 7.812608242034912,
|
| 487 |
+
"logps/chosen": -0.3649575412273407,
|
| 488 |
+
"logps/rejected": -0.8042632937431335,
|
| 489 |
+
"loss": 0.7391,
|
| 490 |
+
"rewards/accuracies": 0.5249999761581421,
|
| 491 |
+
"rewards/chosen": -0.5474363565444946,
|
| 492 |
+
"rewards/margins": 0.6589586734771729,
|
| 493 |
+
"rewards/rejected": -1.206395149230957,
|
| 494 |
+
"step": 270
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 0.29850746268656714,
|
| 498 |
+
"grad_norm": 0.5339816808700562,
|
| 499 |
+
"learning_rate": 4.582303101775249e-06,
|
| 500 |
+
"logits/chosen": 6.687758445739746,
|
| 501 |
+
"logits/rejected": 6.233181476593018,
|
| 502 |
+
"logps/chosen": -0.415935218334198,
|
| 503 |
+
"logps/rejected": -1.2987438440322876,
|
| 504 |
+
"loss": 0.7419,
|
| 505 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 506 |
+
"rewards/chosen": -0.6239027976989746,
|
| 507 |
+
"rewards/margins": 1.3242127895355225,
|
| 508 |
+
"rewards/rejected": -1.9481157064437866,
|
| 509 |
+
"step": 280
|
| 510 |
+
},
|
| 511 |
+
{
|
| 512 |
+
"epoch": 0.3091684434968017,
|
| 513 |
+
"grad_norm": 0.3514000475406647,
|
| 514 |
+
"learning_rate": 4.55287302283426e-06,
|
| 515 |
+
"logits/chosen": 6.2503981590271,
|
| 516 |
+
"logits/rejected": 5.798542499542236,
|
| 517 |
+
"logps/chosen": -0.4319223463535309,
|
| 518 |
+
"logps/rejected": -1.2257453203201294,
|
| 519 |
+
"loss": 0.7235,
|
| 520 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 521 |
+
"rewards/chosen": -0.6478835344314575,
|
| 522 |
+
"rewards/margins": 1.1907342672348022,
|
| 523 |
+
"rewards/rejected": -1.8386180400848389,
|
| 524 |
+
"step": 290
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 0.31982942430703626,
|
| 528 |
+
"grad_norm": 0.6761008501052856,
|
| 529 |
+
"learning_rate": 4.522542485937369e-06,
|
| 530 |
+
"logits/chosen": 4.4480695724487305,
|
| 531 |
+
"logits/rejected": 4.290585994720459,
|
| 532 |
+
"logps/chosen": -0.42002564668655396,
|
| 533 |
+
"logps/rejected": -1.4215493202209473,
|
| 534 |
+
"loss": 0.7058,
|
| 535 |
+
"rewards/accuracies": 0.637499988079071,
|
| 536 |
+
"rewards/chosen": -0.6300384402275085,
|
| 537 |
+
"rewards/margins": 1.5022855997085571,
|
| 538 |
+
"rewards/rejected": -2.132323980331421,
|
| 539 |
+
"step": 300
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"epoch": 0.31982942430703626,
|
| 543 |
+
"eval_logits/chosen": 4.789332389831543,
|
| 544 |
+
"eval_logits/rejected": 4.481485366821289,
|
| 545 |
+
"eval_logps/chosen": -0.4049508571624756,
|
| 546 |
+
"eval_logps/rejected": -1.395646095275879,
|
| 547 |
+
"eval_loss": 0.6695442199707031,
|
| 548 |
+
"eval_rewards/accuracies": 0.6710526347160339,
|
| 549 |
+
"eval_rewards/chosen": -0.6074262857437134,
|
| 550 |
+
"eval_rewards/margins": 1.4860429763793945,
|
| 551 |
+
"eval_rewards/rejected": -2.0934693813323975,
|
| 552 |
+
"eval_runtime": 21.4397,
|
| 553 |
+
"eval_samples_per_second": 28.312,
|
| 554 |
+
"eval_steps_per_second": 3.545,
|
| 555 |
+
"step": 300
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.3304904051172708,
|
| 559 |
+
"grad_norm": 0.44682690501213074,
|
| 560 |
+
"learning_rate": 4.491324795060491e-06,
|
| 561 |
+
"logits/chosen": 5.487166404724121,
|
| 562 |
+
"logits/rejected": 4.501384258270264,
|
| 563 |
+
"logps/chosen": -0.5215579867362976,
|
| 564 |
+
"logps/rejected": -1.7223398685455322,
|
| 565 |
+
"loss": 0.6988,
|
| 566 |
+
"rewards/accuracies": 0.612500011920929,
|
| 567 |
+
"rewards/chosen": -0.7823370695114136,
|
| 568 |
+
"rewards/margins": 1.8011726140975952,
|
| 569 |
+
"rewards/rejected": -2.5835094451904297,
|
| 570 |
+
"step": 310
|
| 571 |
+
},
|
| 572 |
+
{
|
| 573 |
+
"epoch": 0.3411513859275053,
|
| 574 |
+
"grad_norm": 0.41085830330848694,
|
| 575 |
+
"learning_rate": 4.4592336433146e-06,
|
| 576 |
+
"logits/chosen": 4.162590026855469,
|
| 577 |
+
"logits/rejected": 2.876271963119507,
|
| 578 |
+
"logps/chosen": -0.5402930974960327,
|
| 579 |
+
"logps/rejected": -1.7925996780395508,
|
| 580 |
+
"loss": 0.6811,
|
| 581 |
+
"rewards/accuracies": 0.625,
|
| 582 |
+
"rewards/chosen": -0.8104397058486938,
|
| 583 |
+
"rewards/margins": 1.8784599304199219,
|
| 584 |
+
"rewards/rejected": -2.688899517059326,
|
| 585 |
+
"step": 320
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 0.35181236673773986,
|
| 589 |
+
"grad_norm": 0.5611584186553955,
|
| 590 |
+
"learning_rate": 4.426283106939474e-06,
|
| 591 |
+
"logits/chosen": 4.088540077209473,
|
| 592 |
+
"logits/rejected": 3.081679582595825,
|
| 593 |
+
"logps/chosen": -0.541223406791687,
|
| 594 |
+
"logps/rejected": -1.9464069604873657,
|
| 595 |
+
"loss": 0.6614,
|
| 596 |
+
"rewards/accuracies": 0.699999988079071,
|
| 597 |
+
"rewards/chosen": -0.8118351101875305,
|
| 598 |
+
"rewards/margins": 2.1077752113342285,
|
| 599 |
+
"rewards/rejected": -2.919610023498535,
|
| 600 |
+
"step": 330
|
| 601 |
+
},
|
| 602 |
+
{
|
| 603 |
+
"epoch": 0.3624733475479744,
|
| 604 |
+
"grad_norm": 4.05828857421875,
|
| 605 |
+
"learning_rate": 4.3924876391293915e-06,
|
| 606 |
+
"logits/chosen": 3.3937134742736816,
|
| 607 |
+
"logits/rejected": 2.4182538986206055,
|
| 608 |
+
"logps/chosen": -0.6656067967414856,
|
| 609 |
+
"logps/rejected": -1.5255868434906006,
|
| 610 |
+
"loss": 0.6583,
|
| 611 |
+
"rewards/accuracies": 0.574999988079071,
|
| 612 |
+
"rewards/chosen": -0.9984102249145508,
|
| 613 |
+
"rewards/margins": 1.2899701595306396,
|
| 614 |
+
"rewards/rejected": -2.2883803844451904,
|
| 615 |
+
"step": 340
|
| 616 |
+
},
|
| 617 |
+
{
|
| 618 |
+
"epoch": 0.373134328358209,
|
| 619 |
+
"grad_norm": 0.8311880230903625,
|
| 620 |
+
"learning_rate": 4.357862063693486e-06,
|
| 621 |
+
"logits/chosen": 2.503194570541382,
|
| 622 |
+
"logits/rejected": 1.5284960269927979,
|
| 623 |
+
"logps/chosen": -0.6593035459518433,
|
| 624 |
+
"logps/rejected": -2.211193323135376,
|
| 625 |
+
"loss": 0.5911,
|
| 626 |
+
"rewards/accuracies": 0.699999988079071,
|
| 627 |
+
"rewards/chosen": -0.9889553189277649,
|
| 628 |
+
"rewards/margins": 2.3278346061706543,
|
| 629 |
+
"rewards/rejected": -3.3167896270751953,
|
| 630 |
+
"step": 350
|
| 631 |
+
},
|
| 632 |
+
{
|
| 633 |
+
"epoch": 0.373134328358209,
|
| 634 |
+
"eval_logits/chosen": 2.556962728500366,
|
| 635 |
+
"eval_logits/rejected": 1.830418586730957,
|
| 636 |
+
"eval_logps/chosen": -0.6546408534049988,
|
| 637 |
+
"eval_logps/rejected": -1.9014692306518555,
|
| 638 |
+
"eval_loss": 0.5961893200874329,
|
| 639 |
+
"eval_rewards/accuracies": 0.6842105388641357,
|
| 640 |
+
"eval_rewards/chosen": -0.9819613099098206,
|
| 641 |
+
"eval_rewards/margins": 1.8702424764633179,
|
| 642 |
+
"eval_rewards/rejected": -2.852203845977783,
|
| 643 |
+
"eval_runtime": 21.4393,
|
| 644 |
+
"eval_samples_per_second": 28.312,
|
| 645 |
+
"eval_steps_per_second": 3.545,
|
| 646 |
+
"step": 350
|
| 647 |
+
},
|
| 648 |
+
{
|
| 649 |
+
"epoch": 0.3837953091684435,
|
| 650 |
+
"grad_norm": 1.4237236976623535,
|
| 651 |
+
"learning_rate": 4.322421568553529e-06,
|
| 652 |
+
"logits/chosen": 3.0001542568206787,
|
| 653 |
+
"logits/rejected": 1.9715242385864258,
|
| 654 |
+
"logps/chosen": -0.8050466775894165,
|
| 655 |
+
"logps/rejected": -2.2938907146453857,
|
| 656 |
+
"loss": 0.58,
|
| 657 |
+
"rewards/accuracies": 0.737500011920929,
|
| 658 |
+
"rewards/chosen": -1.20756995677948,
|
| 659 |
+
"rewards/margins": 2.2332661151885986,
|
| 660 |
+
"rewards/rejected": -3.440835952758789,
|
| 661 |
+
"step": 360
|
| 662 |
+
},
|
| 663 |
+
{
|
| 664 |
+
"epoch": 0.39445628997867804,
|
| 665 |
+
"grad_norm": 2.2651443481445312,
|
| 666 |
+
"learning_rate": 4.286181699082008e-06,
|
| 667 |
+
"logits/chosen": 2.7526040077209473,
|
| 668 |
+
"logits/rejected": 2.05066180229187,
|
| 669 |
+
"logps/chosen": -1.6301355361938477,
|
| 670 |
+
"logps/rejected": -2.9630703926086426,
|
| 671 |
+
"loss": 0.5823,
|
| 672 |
+
"rewards/accuracies": 0.75,
|
| 673 |
+
"rewards/chosen": -2.4452033042907715,
|
| 674 |
+
"rewards/margins": 1.999402642250061,
|
| 675 |
+
"rewards/rejected": -4.444605827331543,
|
| 676 |
+
"step": 370
|
| 677 |
+
},
|
| 678 |
+
{
|
| 679 |
+
"epoch": 0.4051172707889126,
|
| 680 |
+
"grad_norm": 1.9120367765426636,
|
| 681 |
+
"learning_rate": 4.249158351283414e-06,
|
| 682 |
+
"logits/chosen": 1.9757938385009766,
|
| 683 |
+
"logits/rejected": 1.5915673971176147,
|
| 684 |
+
"logps/chosen": -2.063323497772217,
|
| 685 |
+
"logps/rejected": -2.899749755859375,
|
| 686 |
+
"loss": 0.5675,
|
| 687 |
+
"rewards/accuracies": 0.737500011920929,
|
| 688 |
+
"rewards/chosen": -3.094984769821167,
|
| 689 |
+
"rewards/margins": 1.2546398639678955,
|
| 690 |
+
"rewards/rejected": -4.3496246337890625,
|
| 691 |
+
"step": 380
|
| 692 |
+
},
|
| 693 |
+
{
|
| 694 |
+
"epoch": 0.4157782515991471,
|
| 695 |
+
"grad_norm": 3.0018720626831055,
|
| 696 |
+
"learning_rate": 4.211367764821722e-06,
|
| 697 |
+
"logits/chosen": 2.541440486907959,
|
| 698 |
+
"logits/rejected": 1.7436832189559937,
|
| 699 |
+
"logps/chosen": -2.279510736465454,
|
| 700 |
+
"logps/rejected": -3.3447775840759277,
|
| 701 |
+
"loss": 0.4969,
|
| 702 |
+
"rewards/accuracies": 0.8500000238418579,
|
| 703 |
+
"rewards/chosen": -3.4192657470703125,
|
| 704 |
+
"rewards/margins": 1.5979007482528687,
|
| 705 |
+
"rewards/rejected": -5.017167091369629,
|
| 706 |
+
"step": 390
|
| 707 |
+
},
|
| 708 |
+
{
|
| 709 |
+
"epoch": 0.42643923240938164,
|
| 710 |
+
"grad_norm": 1.9656275510787964,
|
| 711 |
+
"learning_rate": 4.172826515897146e-06,
|
| 712 |
+
"logits/chosen": 1.6748476028442383,
|
| 713 |
+
"logits/rejected": 1.0921740531921387,
|
| 714 |
+
"logps/chosen": -2.147991180419922,
|
| 715 |
+
"logps/rejected": -3.380042314529419,
|
| 716 |
+
"loss": 0.5135,
|
| 717 |
+
"rewards/accuracies": 0.8374999761581421,
|
| 718 |
+
"rewards/chosen": -3.221986770629883,
|
| 719 |
+
"rewards/margins": 1.8480768203735352,
|
| 720 |
+
"rewards/rejected": -5.07006311416626,
|
| 721 |
+
"step": 400
|
| 722 |
+
},
|
| 723 |
+
{
|
| 724 |
+
"epoch": 0.42643923240938164,
|
| 725 |
+
"eval_logits/chosen": 2.210231065750122,
|
| 726 |
+
"eval_logits/rejected": 1.679926872253418,
|
| 727 |
+
"eval_logps/chosen": -2.044506788253784,
|
| 728 |
+
"eval_logps/rejected": -3.713956356048584,
|
| 729 |
+
"eval_loss": 0.47455134987831116,
|
| 730 |
+
"eval_rewards/accuracies": 0.9342105388641357,
|
| 731 |
+
"eval_rewards/chosen": -3.0667598247528076,
|
| 732 |
+
"eval_rewards/margins": 2.5041754245758057,
|
| 733 |
+
"eval_rewards/rejected": -5.570935249328613,
|
| 734 |
+
"eval_runtime": 21.4401,
|
| 735 |
+
"eval_samples_per_second": 28.311,
|
| 736 |
+
"eval_steps_per_second": 3.545,
|
| 737 |
+
"step": 400
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 0.43710021321961623,
|
| 741 |
+
"grad_norm": 2.501361131668091,
|
| 742 |
+
"learning_rate": 4.133551509975264e-06,
|
| 743 |
+
"logits/chosen": 1.9820306301116943,
|
| 744 |
+
"logits/rejected": 1.3992068767547607,
|
| 745 |
+
"logps/chosen": -2.300197124481201,
|
| 746 |
+
"logps/rejected": -3.813164472579956,
|
| 747 |
+
"loss": 0.498,
|
| 748 |
+
"rewards/accuracies": 0.8500000238418579,
|
| 749 |
+
"rewards/chosen": -3.4502956867218018,
|
| 750 |
+
"rewards/margins": 2.2694506645202637,
|
| 751 |
+
"rewards/rejected": -5.7197465896606445,
|
| 752 |
+
"step": 410
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 0.44776119402985076,
|
| 756 |
+
"grad_norm": 3.828648090362549,
|
| 757 |
+
"learning_rate": 4.093559974371725e-06,
|
| 758 |
+
"logits/chosen": 2.7997095584869385,
|
| 759 |
+
"logits/rejected": 2.4387598037719727,
|
| 760 |
+
"logps/chosen": -2.687736749649048,
|
| 761 |
+
"logps/rejected": -4.425741195678711,
|
| 762 |
+
"loss": 0.4494,
|
| 763 |
+
"rewards/accuracies": 0.862500011920929,
|
| 764 |
+
"rewards/chosen": -4.031605243682861,
|
| 765 |
+
"rewards/margins": 2.607006788253784,
|
| 766 |
+
"rewards/rejected": -6.638613224029541,
|
| 767 |
+
"step": 420
|
| 768 |
+
},
|
| 769 |
+
{
|
| 770 |
+
"epoch": 0.4584221748400853,
|
| 771 |
+
"grad_norm": 2.635803461074829,
|
| 772 |
+
"learning_rate": 4.052869450695776e-06,
|
| 773 |
+
"logits/chosen": 2.942661762237549,
|
| 774 |
+
"logits/rejected": 2.019963026046753,
|
| 775 |
+
"logps/chosen": -2.98117733001709,
|
| 776 |
+
"logps/rejected": -4.717232704162598,
|
| 777 |
+
"loss": 0.4796,
|
| 778 |
+
"rewards/accuracies": 0.8500000238418579,
|
| 779 |
+
"rewards/chosen": -4.471765518188477,
|
| 780 |
+
"rewards/margins": 2.60408353805542,
|
| 781 |
+
"rewards/rejected": -7.075850009918213,
|
| 782 |
+
"step": 430
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 0.4690831556503198,
|
| 786 |
+
"grad_norm": 3.140829086303711,
|
| 787 |
+
"learning_rate": 4.011497787155938e-06,
|
| 788 |
+
"logits/chosen": 3.2747459411621094,
|
| 789 |
+
"logits/rejected": 2.2958083152770996,
|
| 790 |
+
"logps/chosen": -3.129321575164795,
|
| 791 |
+
"logps/rejected": -4.921725273132324,
|
| 792 |
+
"loss": 0.4468,
|
| 793 |
+
"rewards/accuracies": 0.862500011920929,
|
| 794 |
+
"rewards/chosen": -4.69398307800293,
|
| 795 |
+
"rewards/margins": 2.688605785369873,
|
| 796 |
+
"rewards/rejected": -7.3825883865356445,
|
| 797 |
+
"step": 440
|
| 798 |
+
},
|
| 799 |
+
{
|
| 800 |
+
"epoch": 0.47974413646055436,
|
| 801 |
+
"grad_norm": 2.7932240962982178,
|
| 802 |
+
"learning_rate": 3.969463130731183e-06,
|
| 803 |
+
"logits/chosen": 2.205420970916748,
|
| 804 |
+
"logits/rejected": 1.4024155139923096,
|
| 805 |
+
"logps/chosen": -2.7564563751220703,
|
| 806 |
+
"logps/rejected": -4.563851356506348,
|
| 807 |
+
"loss": 0.4073,
|
| 808 |
+
"rewards/accuracies": 0.887499988079071,
|
| 809 |
+
"rewards/chosen": -4.1346845626831055,
|
| 810 |
+
"rewards/margins": 2.711092472076416,
|
| 811 |
+
"rewards/rejected": -6.8457770347595215,
|
| 812 |
+
"step": 450
|
| 813 |
+
},
|
| 814 |
+
{
|
| 815 |
+
"epoch": 0.47974413646055436,
|
| 816 |
+
"eval_logits/chosen": 2.0136826038360596,
|
| 817 |
+
"eval_logits/rejected": 1.561701774597168,
|
| 818 |
+
"eval_logps/chosen": -2.7486908435821533,
|
| 819 |
+
"eval_logps/rejected": -4.690793514251709,
|
| 820 |
+
"eval_loss": 0.41499289870262146,
|
| 821 |
+
"eval_rewards/accuracies": 0.9210526347160339,
|
| 822 |
+
"eval_rewards/chosen": -4.123035907745361,
|
| 823 |
+
"eval_rewards/margins": 2.913153648376465,
|
| 824 |
+
"eval_rewards/rejected": -7.036189079284668,
|
| 825 |
+
"eval_runtime": 21.4387,
|
| 826 |
+
"eval_samples_per_second": 28.313,
|
| 827 |
+
"eval_steps_per_second": 3.545,
|
| 828 |
+
"step": 450
|
| 829 |
+
},
|
| 830 |
+
{
|
| 831 |
+
"epoch": 0.4904051172707889,
|
| 832 |
+
"grad_norm": 2.7059199810028076,
|
| 833 |
+
"learning_rate": 3.92678391921108e-06,
|
| 834 |
+
"logits/chosen": 2.257246494293213,
|
| 835 |
+
"logits/rejected": 1.6654322147369385,
|
| 836 |
+
"logps/chosen": -3.389554500579834,
|
| 837 |
+
"logps/rejected": -5.6951165199279785,
|
| 838 |
+
"loss": 0.4004,
|
| 839 |
+
"rewards/accuracies": 0.949999988079071,
|
| 840 |
+
"rewards/chosen": -5.084332466125488,
|
| 841 |
+
"rewards/margins": 3.4583427906036377,
|
| 842 |
+
"rewards/rejected": -8.542675018310547,
|
| 843 |
+
"step": 460
|
| 844 |
+
},
|
| 845 |
+
{
|
| 846 |
+
"epoch": 0.5010660980810234,
|
| 847 |
+
"grad_norm": 2.245579719543457,
|
| 848 |
+
"learning_rate": 3.88347887310836e-06,
|
| 849 |
+
"logits/chosen": 2.3386971950531006,
|
| 850 |
+
"logits/rejected": 2.086036205291748,
|
| 851 |
+
"logps/chosen": -3.2753937244415283,
|
| 852 |
+
"logps/rejected": -5.4362359046936035,
|
| 853 |
+
"loss": 0.3976,
|
| 854 |
+
"rewards/accuracies": 0.875,
|
| 855 |
+
"rewards/chosen": -4.913090705871582,
|
| 856 |
+
"rewards/margins": 3.241262912750244,
|
| 857 |
+
"rewards/rejected": -8.154353141784668,
|
| 858 |
+
"step": 470
|
| 859 |
+
},
|
| 860 |
+
{
|
| 861 |
+
"epoch": 0.511727078891258,
|
| 862 |
+
"grad_norm": 2.8131167888641357,
|
| 863 |
+
"learning_rate": 3.839566987447492e-06,
|
| 864 |
+
"logits/chosen": 1.6951004266738892,
|
| 865 |
+
"logits/rejected": 1.3795586824417114,
|
| 866 |
+
"logps/chosen": -3.2933483123779297,
|
| 867 |
+
"logps/rejected": -5.050060749053955,
|
| 868 |
+
"loss": 0.3982,
|
| 869 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 870 |
+
"rewards/chosen": -4.940022945404053,
|
| 871 |
+
"rewards/margins": 2.635068416595459,
|
| 872 |
+
"rewards/rejected": -7.5750908851623535,
|
| 873 |
+
"step": 480
|
| 874 |
+
},
|
| 875 |
+
{
|
| 876 |
+
"epoch": 0.5223880597014925,
|
| 877 |
+
"grad_norm": 2.6465814113616943,
|
| 878 |
+
"learning_rate": 3.795067523432826e-06,
|
| 879 |
+
"logits/chosen": 3.136894702911377,
|
| 880 |
+
"logits/rejected": 2.6332411766052246,
|
| 881 |
+
"logps/chosen": -4.419378757476807,
|
| 882 |
+
"logps/rejected": -6.467301845550537,
|
| 883 |
+
"loss": 0.3974,
|
| 884 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 885 |
+
"rewards/chosen": -6.629067420959473,
|
| 886 |
+
"rewards/margins": 3.071885585784912,
|
| 887 |
+
"rewards/rejected": -9.700952529907227,
|
| 888 |
+
"step": 490
|
| 889 |
+
},
|
| 890 |
+
{
|
| 891 |
+
"epoch": 0.5330490405117271,
|
| 892 |
+
"grad_norm": 3.6718053817749023,
|
| 893 |
+
"learning_rate": 3.7500000000000005e-06,
|
| 894 |
+
"logits/chosen": 1.681780457496643,
|
| 895 |
+
"logits/rejected": 1.0038775205612183,
|
| 896 |
+
"logps/chosen": -3.266970157623291,
|
| 897 |
+
"logps/rejected": -5.594450950622559,
|
| 898 |
+
"loss": 0.367,
|
| 899 |
+
"rewards/accuracies": 0.925000011920929,
|
| 900 |
+
"rewards/chosen": -4.900455474853516,
|
| 901 |
+
"rewards/margins": 3.4912209510803223,
|
| 902 |
+
"rewards/rejected": -8.391676902770996,
|
| 903 |
+
"step": 500
|
| 904 |
+
},
|
| 905 |
+
{
|
| 906 |
+
"epoch": 0.5330490405117271,
|
| 907 |
+
"eval_logits/chosen": 2.110192060470581,
|
| 908 |
+
"eval_logits/rejected": 1.7233155965805054,
|
| 909 |
+
"eval_logps/chosen": -3.0329930782318115,
|
| 910 |
+
"eval_logps/rejected": -5.3280930519104,
|
| 911 |
+
"eval_loss": 0.387028306722641,
|
| 912 |
+
"eval_rewards/accuracies": 0.9210526347160339,
|
| 913 |
+
"eval_rewards/chosen": -4.549489498138428,
|
| 914 |
+
"eval_rewards/margins": 3.4426498413085938,
|
| 915 |
+
"eval_rewards/rejected": -7.9921393394470215,
|
| 916 |
+
"eval_runtime": 21.4417,
|
| 917 |
+
"eval_samples_per_second": 28.309,
|
| 918 |
+
"eval_steps_per_second": 3.545,
|
| 919 |
+
"step": 500
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"epoch": 0.5437100213219617,
|
| 923 |
+
"grad_norm": 3.392418622970581,
|
| 924 |
+
"learning_rate": 3.7043841852542884e-06,
|
| 925 |
+
"logits/chosen": 3.0106852054595947,
|
| 926 |
+
"logits/rejected": 2.309483528137207,
|
| 927 |
+
"logps/chosen": -3.4964828491210938,
|
| 928 |
+
"logps/rejected": -5.438345909118652,
|
| 929 |
+
"loss": 0.3776,
|
| 930 |
+
"rewards/accuracies": 0.925000011920929,
|
| 931 |
+
"rewards/chosen": -5.244723796844482,
|
| 932 |
+
"rewards/margins": 2.9127936363220215,
|
| 933 |
+
"rewards/rejected": -8.15751838684082,
|
| 934 |
+
"step": 510
|
| 935 |
+
},
|
| 936 |
+
{
|
| 937 |
+
"epoch": 0.5543710021321961,
|
| 938 |
+
"grad_norm": 3.551785707473755,
|
| 939 |
+
"learning_rate": 3.658240087799655e-06,
|
| 940 |
+
"logits/chosen": 2.8624072074890137,
|
| 941 |
+
"logits/rejected": 2.425307512283325,
|
| 942 |
+
"logps/chosen": -3.780029773712158,
|
| 943 |
+
"logps/rejected": -5.929925918579102,
|
| 944 |
+
"loss": 0.3999,
|
| 945 |
+
"rewards/accuracies": 0.887499988079071,
|
| 946 |
+
"rewards/chosen": -5.670044898986816,
|
| 947 |
+
"rewards/margins": 3.224843978881836,
|
| 948 |
+
"rewards/rejected": -8.894887924194336,
|
| 949 |
+
"step": 520
|
| 950 |
+
},
|
| 951 |
+
{
|
| 952 |
+
"epoch": 0.5650319829424307,
|
| 953 |
+
"grad_norm": 3.0529866218566895,
|
| 954 |
+
"learning_rate": 3.611587947962319e-06,
|
| 955 |
+
"logits/chosen": 2.3145668506622314,
|
| 956 |
+
"logits/rejected": 1.5088326930999756,
|
| 957 |
+
"logps/chosen": -3.1259171962738037,
|
| 958 |
+
"logps/rejected": -5.282050132751465,
|
| 959 |
+
"loss": 0.3346,
|
| 960 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 961 |
+
"rewards/chosen": -4.688876152038574,
|
| 962 |
+
"rewards/margins": 3.234198808670044,
|
| 963 |
+
"rewards/rejected": -7.923074245452881,
|
| 964 |
+
"step": 530
|
| 965 |
+
},
|
| 966 |
+
{
|
| 967 |
+
"epoch": 0.5756929637526652,
|
| 968 |
+
"grad_norm": 1.7146470546722412,
|
| 969 |
+
"learning_rate": 3.564448228912682e-06,
|
| 970 |
+
"logits/chosen": 2.5195610523223877,
|
| 971 |
+
"logits/rejected": 2.286477565765381,
|
| 972 |
+
"logps/chosen": -4.000255107879639,
|
| 973 |
+
"logps/rejected": -6.475634574890137,
|
| 974 |
+
"loss": 0.3675,
|
| 975 |
+
"rewards/accuracies": 0.925000011920929,
|
| 976 |
+
"rewards/chosen": -6.000383377075195,
|
| 977 |
+
"rewards/margins": 3.713069438934326,
|
| 978 |
+
"rewards/rejected": -9.713452339172363,
|
| 979 |
+
"step": 540
|
| 980 |
+
},
|
| 981 |
+
{
|
| 982 |
+
"epoch": 0.5863539445628998,
|
| 983 |
+
"grad_norm": 8.31704330444336,
|
| 984 |
+
"learning_rate": 3.516841607689501e-06,
|
| 985 |
+
"logits/chosen": 1.8846759796142578,
|
| 986 |
+
"logits/rejected": 1.5663390159606934,
|
| 987 |
+
"logps/chosen": -3.9287109375,
|
| 988 |
+
"logps/rejected": -6.041600704193115,
|
| 989 |
+
"loss": 0.3921,
|
| 990 |
+
"rewards/accuracies": 0.862500011920929,
|
| 991 |
+
"rewards/chosen": -5.89306640625,
|
| 992 |
+
"rewards/margins": 3.1693339347839355,
|
| 993 |
+
"rewards/rejected": -9.062400817871094,
|
| 994 |
+
"step": 550
|
| 995 |
+
},
|
| 996 |
+
{
|
| 997 |
+
"epoch": 0.5863539445628998,
|
| 998 |
+
"eval_logits/chosen": 2.0545763969421387,
|
| 999 |
+
"eval_logits/rejected": 1.7521167993545532,
|
| 1000 |
+
"eval_logps/chosen": -3.2991809844970703,
|
| 1001 |
+
"eval_logps/rejected": -5.87559175491333,
|
| 1002 |
+
"eval_loss": 0.35785165429115295,
|
| 1003 |
+
"eval_rewards/accuracies": 0.9342105388641357,
|
| 1004 |
+
"eval_rewards/chosen": -4.948771953582764,
|
| 1005 |
+
"eval_rewards/margins": 3.8646163940429688,
|
| 1006 |
+
"eval_rewards/rejected": -8.813387870788574,
|
| 1007 |
+
"eval_runtime": 21.439,
|
| 1008 |
+
"eval_samples_per_second": 28.313,
|
| 1009 |
+
"eval_steps_per_second": 3.545,
|
| 1010 |
+
"step": 550
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 0.5970149253731343,
|
| 1014 |
+
"grad_norm": 4.491275787353516,
|
| 1015 |
+
"learning_rate": 3.4687889661302577e-06,
|
| 1016 |
+
"logits/chosen": 2.9886631965637207,
|
| 1017 |
+
"logits/rejected": 2.2341508865356445,
|
| 1018 |
+
"logps/chosen": -4.050547122955322,
|
| 1019 |
+
"logps/rejected": -6.301305294036865,
|
| 1020 |
+
"loss": 0.3701,
|
| 1021 |
+
"rewards/accuracies": 0.887499988079071,
|
| 1022 |
+
"rewards/chosen": -6.0758209228515625,
|
| 1023 |
+
"rewards/margins": 3.3761372566223145,
|
| 1024 |
+
"rewards/rejected": -9.451958656311035,
|
| 1025 |
+
"step": 560
|
| 1026 |
+
},
|
| 1027 |
+
{
|
| 1028 |
+
"epoch": 0.6076759061833689,
|
| 1029 |
+
"grad_norm": 2.40407395362854,
|
| 1030 |
+
"learning_rate": 3.4203113817116955e-06,
|
| 1031 |
+
"logits/chosen": 3.0933375358581543,
|
| 1032 |
+
"logits/rejected": 2.4652345180511475,
|
| 1033 |
+
"logps/chosen": -3.9168992042541504,
|
| 1034 |
+
"logps/rejected": -5.933487892150879,
|
| 1035 |
+
"loss": 0.3481,
|
| 1036 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 1037 |
+
"rewards/chosen": -5.875349521636963,
|
| 1038 |
+
"rewards/margins": 3.024883508682251,
|
| 1039 |
+
"rewards/rejected": -8.900232315063477,
|
| 1040 |
+
"step": 570
|
| 1041 |
+
},
|
| 1042 |
+
{
|
| 1043 |
+
"epoch": 0.6183368869936035,
|
| 1044 |
+
"grad_norm": 4.073615550994873,
|
| 1045 |
+
"learning_rate": 3.3714301183045382e-06,
|
| 1046 |
+
"logits/chosen": 2.5948500633239746,
|
| 1047 |
+
"logits/rejected": 2.202971935272217,
|
| 1048 |
+
"logps/chosen": -3.6740527153015137,
|
| 1049 |
+
"logps/rejected": -6.205447196960449,
|
| 1050 |
+
"loss": 0.3571,
|
| 1051 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 1052 |
+
"rewards/chosen": -5.511078834533691,
|
| 1053 |
+
"rewards/margins": 3.7970924377441406,
|
| 1054 |
+
"rewards/rejected": -9.308171272277832,
|
| 1055 |
+
"step": 580
|
| 1056 |
+
},
|
| 1057 |
+
{
|
| 1058 |
+
"epoch": 0.6289978678038379,
|
| 1059 |
+
"grad_norm": 2.427555799484253,
|
| 1060 |
+
"learning_rate": 3.3221666168464584e-06,
|
| 1061 |
+
"logits/chosen": 1.6076780557632446,
|
| 1062 |
+
"logits/rejected": 1.0556285381317139,
|
| 1063 |
+
"logps/chosen": -3.609313488006592,
|
| 1064 |
+
"logps/rejected": -5.912892818450928,
|
| 1065 |
+
"loss": 0.3666,
|
| 1066 |
+
"rewards/accuracies": 0.875,
|
| 1067 |
+
"rewards/chosen": -5.413969993591309,
|
| 1068 |
+
"rewards/margins": 3.455368757247925,
|
| 1069 |
+
"rewards/rejected": -8.869338989257812,
|
| 1070 |
+
"step": 590
|
| 1071 |
+
},
|
| 1072 |
+
{
|
| 1073 |
+
"epoch": 0.6396588486140725,
|
| 1074 |
+
"grad_norm": 3.910998821258545,
|
| 1075 |
+
"learning_rate": 3.272542485937369e-06,
|
| 1076 |
+
"logits/chosen": 2.5494394302368164,
|
| 1077 |
+
"logits/rejected": 1.9157211780548096,
|
| 1078 |
+
"logps/chosen": -4.038924217224121,
|
| 1079 |
+
"logps/rejected": -6.580315589904785,
|
| 1080 |
+
"loss": 0.3017,
|
| 1081 |
+
"rewards/accuracies": 0.862500011920929,
|
| 1082 |
+
"rewards/chosen": -6.058385372161865,
|
| 1083 |
+
"rewards/margins": 3.812087297439575,
|
| 1084 |
+
"rewards/rejected": -9.87047290802002,
|
| 1085 |
+
"step": 600
|
| 1086 |
+
},
|
| 1087 |
+
{
|
| 1088 |
+
"epoch": 0.6396588486140725,
|
| 1089 |
+
"eval_logits/chosen": 2.179999589920044,
|
| 1090 |
+
"eval_logits/rejected": 1.918432593345642,
|
| 1091 |
+
"eval_logps/chosen": -3.3114354610443115,
|
| 1092 |
+
"eval_logps/rejected": -6.157403469085693,
|
| 1093 |
+
"eval_loss": 0.3409230411052704,
|
| 1094 |
+
"eval_rewards/accuracies": 0.9210526347160339,
|
| 1095 |
+
"eval_rewards/chosen": -4.967153549194336,
|
| 1096 |
+
"eval_rewards/margins": 4.268952369689941,
|
| 1097 |
+
"eval_rewards/rejected": -9.236105918884277,
|
| 1098 |
+
"eval_runtime": 21.4405,
|
| 1099 |
+
"eval_samples_per_second": 28.311,
|
| 1100 |
+
"eval_steps_per_second": 3.545,
|
| 1101 |
+
"step": 600
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"epoch": 0.650319829424307,
|
| 1105 |
+
"grad_norm": 3.6193580627441406,
|
| 1106 |
+
"learning_rate": 3.222579492361179e-06,
|
| 1107 |
+
"logits/chosen": 2.680032968521118,
|
| 1108 |
+
"logits/rejected": 1.9571218490600586,
|
| 1109 |
+
"logps/chosen": -3.6821212768554688,
|
| 1110 |
+
"logps/rejected": -6.76863956451416,
|
| 1111 |
+
"loss": 0.3429,
|
| 1112 |
+
"rewards/accuracies": 0.9375,
|
| 1113 |
+
"rewards/chosen": -5.523181915283203,
|
| 1114 |
+
"rewards/margins": 4.629776954650879,
|
| 1115 |
+
"rewards/rejected": -10.152959823608398,
|
| 1116 |
+
"step": 610
|
| 1117 |
+
},
|
| 1118 |
+
{
|
| 1119 |
+
"epoch": 0.6609808102345416,
|
| 1120 |
+
"grad_norm": 4.7948808670043945,
|
| 1121 |
+
"learning_rate": 3.1722995515381644e-06,
|
| 1122 |
+
"logits/chosen": 2.570702075958252,
|
| 1123 |
+
"logits/rejected": 2.2683706283569336,
|
| 1124 |
+
"logps/chosen": -4.124785423278809,
|
| 1125 |
+
"logps/rejected": -6.595047950744629,
|
| 1126 |
+
"loss": 0.3355,
|
| 1127 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1128 |
+
"rewards/chosen": -6.187178134918213,
|
| 1129 |
+
"rewards/margins": 3.7053933143615723,
|
| 1130 |
+
"rewards/rejected": -9.892572402954102,
|
| 1131 |
+
"step": 620
|
| 1132 |
+
},
|
| 1133 |
+
{
|
| 1134 |
+
"epoch": 0.6716417910447762,
|
| 1135 |
+
"grad_norm": 2.727231025695801,
|
| 1136 |
+
"learning_rate": 3.121724717912138e-06,
|
| 1137 |
+
"logits/chosen": 2.0630054473876953,
|
| 1138 |
+
"logits/rejected": 1.4909979104995728,
|
| 1139 |
+
"logps/chosen": -3.6443302631378174,
|
| 1140 |
+
"logps/rejected": -6.239049911499023,
|
| 1141 |
+
"loss": 0.3218,
|
| 1142 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1143 |
+
"rewards/chosen": -5.466495513916016,
|
| 1144 |
+
"rewards/margins": 3.8920791149139404,
|
| 1145 |
+
"rewards/rejected": -9.358574867248535,
|
| 1146 |
+
"step": 630
|
| 1147 |
+
},
|
| 1148 |
+
{
|
| 1149 |
+
"epoch": 0.6823027718550106,
|
| 1150 |
+
"grad_norm": 4.138846397399902,
|
| 1151 |
+
"learning_rate": 3.0708771752766397e-06,
|
| 1152 |
+
"logits/chosen": 1.6833502054214478,
|
| 1153 |
+
"logits/rejected": 1.2245066165924072,
|
| 1154 |
+
"logps/chosen": -4.318512916564941,
|
| 1155 |
+
"logps/rejected": -7.222353935241699,
|
| 1156 |
+
"loss": 0.2868,
|
| 1157 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1158 |
+
"rewards/chosen": -6.477769374847412,
|
| 1159 |
+
"rewards/margins": 4.35576057434082,
|
| 1160 |
+
"rewards/rejected": -10.833529472351074,
|
| 1161 |
+
"step": 640
|
| 1162 |
+
},
|
| 1163 |
+
{
|
| 1164 |
+
"epoch": 0.6929637526652452,
|
| 1165 |
+
"grad_norm": 7.172155380249023,
|
| 1166 |
+
"learning_rate": 3.019779227044398e-06,
|
| 1167 |
+
"logits/chosen": 2.5162904262542725,
|
| 1168 |
+
"logits/rejected": 2.0894787311553955,
|
| 1169 |
+
"logps/chosen": -4.13573694229126,
|
| 1170 |
+
"logps/rejected": -7.199030876159668,
|
| 1171 |
+
"loss": 0.3491,
|
| 1172 |
+
"rewards/accuracies": 0.9375,
|
| 1173 |
+
"rewards/chosen": -6.2036051750183105,
|
| 1174 |
+
"rewards/margins": 4.594940185546875,
|
| 1175 |
+
"rewards/rejected": -10.798544883728027,
|
| 1176 |
+
"step": 650
|
| 1177 |
+
},
|
| 1178 |
+
{
|
| 1179 |
+
"epoch": 0.6929637526652452,
|
| 1180 |
+
"eval_logits/chosen": 2.003471612930298,
|
| 1181 |
+
"eval_logits/rejected": 1.8008451461791992,
|
| 1182 |
+
"eval_logps/chosen": -3.2873525619506836,
|
| 1183 |
+
"eval_logps/rejected": -6.338763236999512,
|
| 1184 |
+
"eval_loss": 0.31618499755859375,
|
| 1185 |
+
"eval_rewards/accuracies": 0.9342105388641357,
|
| 1186 |
+
"eval_rewards/chosen": -4.931028842926025,
|
| 1187 |
+
"eval_rewards/margins": 4.5771164894104,
|
| 1188 |
+
"eval_rewards/rejected": -9.508145332336426,
|
| 1189 |
+
"eval_runtime": 21.415,
|
| 1190 |
+
"eval_samples_per_second": 28.345,
|
| 1191 |
+
"eval_steps_per_second": 3.549,
|
| 1192 |
+
"step": 650
|
| 1193 |
+
},
|
| 1194 |
+
{
|
| 1195 |
+
"epoch": 0.7036247334754797,
|
| 1196 |
+
"grad_norm": 3.3068535327911377,
|
| 1197 |
+
"learning_rate": 2.9684532864643123e-06,
|
| 1198 |
+
"logits/chosen": 2.852752208709717,
|
| 1199 |
+
"logits/rejected": 2.3139753341674805,
|
| 1200 |
+
"logps/chosen": -3.9495348930358887,
|
| 1201 |
+
"logps/rejected": -6.509753227233887,
|
| 1202 |
+
"loss": 0.2684,
|
| 1203 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 1204 |
+
"rewards/chosen": -5.924302101135254,
|
| 1205 |
+
"rewards/margins": 3.8403282165527344,
|
| 1206 |
+
"rewards/rejected": -9.764630317687988,
|
| 1207 |
+
"step": 660
|
| 1208 |
+
},
|
| 1209 |
+
{
|
| 1210 |
+
"epoch": 0.7142857142857143,
|
| 1211 |
+
"grad_norm": 2.737560987472534,
|
| 1212 |
+
"learning_rate": 2.9169218667902562e-06,
|
| 1213 |
+
"logits/chosen": 3.034759998321533,
|
| 1214 |
+
"logits/rejected": 2.2936713695526123,
|
| 1215 |
+
"logps/chosen": -3.855508327484131,
|
| 1216 |
+
"logps/rejected": -6.571882724761963,
|
| 1217 |
+
"loss": 0.2887,
|
| 1218 |
+
"rewards/accuracies": 0.9375,
|
| 1219 |
+
"rewards/chosen": -5.783262252807617,
|
| 1220 |
+
"rewards/margins": 4.0745625495910645,
|
| 1221 |
+
"rewards/rejected": -9.857824325561523,
|
| 1222 |
+
"step": 670
|
| 1223 |
+
},
|
| 1224 |
+
{
|
| 1225 |
+
"epoch": 0.7249466950959488,
|
| 1226 |
+
"grad_norm": 4.387801647186279,
|
| 1227 |
+
"learning_rate": 2.8652075714060296e-06,
|
| 1228 |
+
"logits/chosen": 1.5638288259506226,
|
| 1229 |
+
"logits/rejected": 1.1110167503356934,
|
| 1230 |
+
"logps/chosen": -3.7569847106933594,
|
| 1231 |
+
"logps/rejected": -6.927552223205566,
|
| 1232 |
+
"loss": 0.3319,
|
| 1233 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 1234 |
+
"rewards/chosen": -5.635476589202881,
|
| 1235 |
+
"rewards/margins": 4.755851745605469,
|
| 1236 |
+
"rewards/rejected": -10.391328811645508,
|
| 1237 |
+
"step": 680
|
| 1238 |
+
},
|
| 1239 |
+
{
|
| 1240 |
+
"epoch": 0.7356076759061834,
|
| 1241 |
+
"grad_norm": 2.357923746109009,
|
| 1242 |
+
"learning_rate": 2.813333083910761e-06,
|
| 1243 |
+
"logits/chosen": 2.058987855911255,
|
| 1244 |
+
"logits/rejected": 1.6591012477874756,
|
| 1245 |
+
"logps/chosen": -4.124215126037598,
|
| 1246 |
+
"logps/rejected": -7.349566459655762,
|
| 1247 |
+
"loss": 0.2362,
|
| 1248 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1249 |
+
"rewards/chosen": -6.186322212219238,
|
| 1250 |
+
"rewards/margins": 4.838027000427246,
|
| 1251 |
+
"rewards/rejected": -11.024351119995117,
|
| 1252 |
+
"step": 690
|
| 1253 |
+
},
|
| 1254 |
+
{
|
| 1255 |
+
"epoch": 0.746268656716418,
|
| 1256 |
+
"grad_norm": 4.507752895355225,
|
| 1257 |
+
"learning_rate": 2.761321158169134e-06,
|
| 1258 |
+
"logits/chosen": 2.295815944671631,
|
| 1259 |
+
"logits/rejected": 1.4525251388549805,
|
| 1260 |
+
"logps/chosen": -3.4264607429504395,
|
| 1261 |
+
"logps/rejected": -6.397761344909668,
|
| 1262 |
+
"loss": 0.3257,
|
| 1263 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1264 |
+
"rewards/chosen": -5.139691352844238,
|
| 1265 |
+
"rewards/margins": 4.45695161819458,
|
| 1266 |
+
"rewards/rejected": -9.596643447875977,
|
| 1267 |
+
"step": 700
|
| 1268 |
+
},
|
| 1269 |
+
{
|
| 1270 |
+
"epoch": 0.746268656716418,
|
| 1271 |
+
"eval_logits/chosen": 1.9169458150863647,
|
| 1272 |
+
"eval_logits/rejected": 1.761539101600647,
|
| 1273 |
+
"eval_logps/chosen": -3.324993848800659,
|
| 1274 |
+
"eval_logps/rejected": -6.471699237823486,
|
| 1275 |
+
"eval_loss": 0.3062504529953003,
|
| 1276 |
+
"eval_rewards/accuracies": 0.9473684430122375,
|
| 1277 |
+
"eval_rewards/chosen": -4.987491607666016,
|
| 1278 |
+
"eval_rewards/margins": 4.720058441162109,
|
| 1279 |
+
"eval_rewards/rejected": -9.707548141479492,
|
| 1280 |
+
"eval_runtime": 21.4456,
|
| 1281 |
+
"eval_samples_per_second": 28.304,
|
| 1282 |
+
"eval_steps_per_second": 3.544,
|
| 1283 |
+
"step": 700
|
| 1284 |
+
},
|
| 1285 |
+
{
|
| 1286 |
+
"epoch": 0.7569296375266524,
|
| 1287 |
+
"grad_norm": 6.864930629730225,
|
| 1288 |
+
"learning_rate": 2.70919460833079e-06,
|
| 1289 |
+
"logits/chosen": 1.8696104288101196,
|
| 1290 |
+
"logits/rejected": 1.4743572473526,
|
| 1291 |
+
"logps/chosen": -3.4536032676696777,
|
| 1292 |
+
"logps/rejected": -6.400083065032959,
|
| 1293 |
+
"loss": 0.3255,
|
| 1294 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 1295 |
+
"rewards/chosen": -5.180405616760254,
|
| 1296 |
+
"rewards/margins": 4.4197187423706055,
|
| 1297 |
+
"rewards/rejected": -9.60012435913086,
|
| 1298 |
+
"step": 710
|
| 1299 |
+
},
|
| 1300 |
+
{
|
| 1301 |
+
"epoch": 0.767590618336887,
|
| 1302 |
+
"grad_norm": 5.012319087982178,
|
| 1303 |
+
"learning_rate": 2.6569762988232838e-06,
|
| 1304 |
+
"logits/chosen": 2.861811399459839,
|
| 1305 |
+
"logits/rejected": 2.266331911087036,
|
| 1306 |
+
"logps/chosen": -4.254661560058594,
|
| 1307 |
+
"logps/rejected": -6.501154899597168,
|
| 1308 |
+
"loss": 0.2845,
|
| 1309 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 1310 |
+
"rewards/chosen": -6.381992816925049,
|
| 1311 |
+
"rewards/margins": 3.3697407245635986,
|
| 1312 |
+
"rewards/rejected": -9.751731872558594,
|
| 1313 |
+
"step": 720
|
| 1314 |
+
},
|
| 1315 |
+
{
|
| 1316 |
+
"epoch": 0.7782515991471215,
|
| 1317 |
+
"grad_norm": 1.460620403289795,
|
| 1318 |
+
"learning_rate": 2.604689134322999e-06,
|
| 1319 |
+
"logits/chosen": 2.917189836502075,
|
| 1320 |
+
"logits/rejected": 2.3850274085998535,
|
| 1321 |
+
"logps/chosen": -3.6072356700897217,
|
| 1322 |
+
"logps/rejected": -6.3972039222717285,
|
| 1323 |
+
"loss": 0.2696,
|
| 1324 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1325 |
+
"rewards/chosen": -5.410854339599609,
|
| 1326 |
+
"rewards/margins": 4.184950828552246,
|
| 1327 |
+
"rewards/rejected": -9.595805168151855,
|
| 1328 |
+
"step": 730
|
| 1329 |
+
},
|
| 1330 |
+
{
|
| 1331 |
+
"epoch": 0.7889125799573561,
|
| 1332 |
+
"grad_norm": 5.478528022766113,
|
| 1333 |
+
"learning_rate": 2.5523560497083927e-06,
|
| 1334 |
+
"logits/chosen": 2.2212367057800293,
|
| 1335 |
+
"logits/rejected": 1.6661183834075928,
|
| 1336 |
+
"logps/chosen": -3.353309154510498,
|
| 1337 |
+
"logps/rejected": -6.527769565582275,
|
| 1338 |
+
"loss": 0.2622,
|
| 1339 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1340 |
+
"rewards/chosen": -5.029964447021484,
|
| 1341 |
+
"rewards/margins": 4.761690616607666,
|
| 1342 |
+
"rewards/rejected": -9.791654586791992,
|
| 1343 |
+
"step": 740
|
| 1344 |
+
},
|
| 1345 |
+
{
|
| 1346 |
+
"epoch": 0.7995735607675906,
|
| 1347 |
+
"grad_norm": 2.7094156742095947,
|
| 1348 |
+
"learning_rate": 2.5e-06,
|
| 1349 |
+
"logits/chosen": 2.9158451557159424,
|
| 1350 |
+
"logits/rejected": 2.286925792694092,
|
| 1351 |
+
"logps/chosen": -4.0971856117248535,
|
| 1352 |
+
"logps/rejected": -6.912562370300293,
|
| 1353 |
+
"loss": 0.3243,
|
| 1354 |
+
"rewards/accuracies": 0.887499988079071,
|
| 1355 |
+
"rewards/chosen": -6.145778179168701,
|
| 1356 |
+
"rewards/margins": 4.2230658531188965,
|
| 1357 |
+
"rewards/rejected": -10.368844985961914,
|
| 1358 |
+
"step": 750
|
| 1359 |
+
},
|
| 1360 |
+
{
|
| 1361 |
+
"epoch": 0.7995735607675906,
|
| 1362 |
+
"eval_logits/chosen": 2.0261385440826416,
|
| 1363 |
+
"eval_logits/rejected": 1.8928862810134888,
|
| 1364 |
+
"eval_logps/chosen": -3.538414478302002,
|
| 1365 |
+
"eval_logps/rejected": -6.846959114074707,
|
| 1366 |
+
"eval_loss": 0.2867603600025177,
|
| 1367 |
+
"eval_rewards/accuracies": 0.9342105388641357,
|
| 1368 |
+
"eval_rewards/chosen": -5.307621479034424,
|
| 1369 |
+
"eval_rewards/margins": 4.962815761566162,
|
| 1370 |
+
"eval_rewards/rejected": -10.270438194274902,
|
| 1371 |
+
"eval_runtime": 21.4446,
|
| 1372 |
+
"eval_samples_per_second": 28.305,
|
| 1373 |
+
"eval_steps_per_second": 3.544,
|
| 1374 |
+
"step": 750
|
| 1375 |
+
},
|
| 1376 |
+
{
|
| 1377 |
+
"epoch": 0.8102345415778252,
|
| 1378 |
+
"grad_norm": 2.477257490158081,
|
| 1379 |
+
"learning_rate": 2.447643950291608e-06,
|
| 1380 |
+
"logits/chosen": 1.8020350933074951,
|
| 1381 |
+
"logits/rejected": 1.3997737169265747,
|
| 1382 |
+
"logps/chosen": -3.34975004196167,
|
| 1383 |
+
"logps/rejected": -6.07710075378418,
|
| 1384 |
+
"loss": 0.2394,
|
| 1385 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 1386 |
+
"rewards/chosen": -5.024625301361084,
|
| 1387 |
+
"rewards/margins": 4.091025352478027,
|
| 1388 |
+
"rewards/rejected": -9.11565113067627,
|
| 1389 |
+
"step": 760
|
| 1390 |
+
},
|
| 1391 |
+
{
|
| 1392 |
+
"epoch": 0.8208955223880597,
|
| 1393 |
+
"grad_norm": 4.726482391357422,
|
| 1394 |
+
"learning_rate": 2.3953108656770018e-06,
|
| 1395 |
+
"logits/chosen": 2.757497787475586,
|
| 1396 |
+
"logits/rejected": 2.435739517211914,
|
| 1397 |
+
"logps/chosen": -4.294445991516113,
|
| 1398 |
+
"logps/rejected": -7.038791656494141,
|
| 1399 |
+
"loss": 0.3113,
|
| 1400 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1401 |
+
"rewards/chosen": -6.441669464111328,
|
| 1402 |
+
"rewards/margins": 4.116518497467041,
|
| 1403 |
+
"rewards/rejected": -10.558187484741211,
|
| 1404 |
+
"step": 770
|
| 1405 |
+
},
|
| 1406 |
+
{
|
| 1407 |
+
"epoch": 0.8315565031982942,
|
| 1408 |
+
"grad_norm": 2.759181022644043,
|
| 1409 |
+
"learning_rate": 2.3430237011767166e-06,
|
| 1410 |
+
"logits/chosen": 2.788053274154663,
|
| 1411 |
+
"logits/rejected": 2.1961398124694824,
|
| 1412 |
+
"logps/chosen": -3.792672634124756,
|
| 1413 |
+
"logps/rejected": -7.1101579666137695,
|
| 1414 |
+
"loss": 0.3043,
|
| 1415 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 1416 |
+
"rewards/chosen": -5.689009666442871,
|
| 1417 |
+
"rewards/margins": 4.976227760314941,
|
| 1418 |
+
"rewards/rejected": -10.665237426757812,
|
| 1419 |
+
"step": 780
|
| 1420 |
+
},
|
| 1421 |
+
{
|
| 1422 |
+
"epoch": 0.8422174840085288,
|
| 1423 |
+
"grad_norm": 3.382880687713623,
|
| 1424 |
+
"learning_rate": 2.290805391669212e-06,
|
| 1425 |
+
"logits/chosen": 2.3644931316375732,
|
| 1426 |
+
"logits/rejected": 1.838971495628357,
|
| 1427 |
+
"logps/chosen": -3.919532060623169,
|
| 1428 |
+
"logps/rejected": -7.125067710876465,
|
| 1429 |
+
"loss": 0.2825,
|
| 1430 |
+
"rewards/accuracies": 0.9375,
|
| 1431 |
+
"rewards/chosen": -5.879298210144043,
|
| 1432 |
+
"rewards/margins": 4.8083038330078125,
|
| 1433 |
+
"rewards/rejected": -10.687601089477539,
|
| 1434 |
+
"step": 790
|
| 1435 |
+
},
|
| 1436 |
+
{
|
| 1437 |
+
"epoch": 0.8528784648187633,
|
| 1438 |
+
"grad_norm": 3.331052541732788,
|
| 1439 |
+
"learning_rate": 2.238678841830867e-06,
|
| 1440 |
+
"logits/chosen": 2.3921093940734863,
|
| 1441 |
+
"logits/rejected": 1.9997615814208984,
|
| 1442 |
+
"logps/chosen": -3.46040678024292,
|
| 1443 |
+
"logps/rejected": -6.8662567138671875,
|
| 1444 |
+
"loss": 0.3094,
|
| 1445 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1446 |
+
"rewards/chosen": -5.190610885620117,
|
| 1447 |
+
"rewards/margins": 5.108774662017822,
|
| 1448 |
+
"rewards/rejected": -10.299385070800781,
|
| 1449 |
+
"step": 800
|
| 1450 |
+
},
|
| 1451 |
+
{
|
| 1452 |
+
"epoch": 0.8528784648187633,
|
| 1453 |
+
"eval_logits/chosen": 2.198673963546753,
|
| 1454 |
+
"eval_logits/rejected": 2.091115713119507,
|
| 1455 |
+
"eval_logps/chosen": -3.606259346008301,
|
| 1456 |
+
"eval_logps/rejected": -7.107792854309082,
|
| 1457 |
+
"eval_loss": 0.27349653840065,
|
| 1458 |
+
"eval_rewards/accuracies": 0.9473684430122375,
|
| 1459 |
+
"eval_rewards/chosen": -5.409388542175293,
|
| 1460 |
+
"eval_rewards/margins": 5.25230073928833,
|
| 1461 |
+
"eval_rewards/rejected": -10.661689758300781,
|
| 1462 |
+
"eval_runtime": 21.4427,
|
| 1463 |
+
"eval_samples_per_second": 28.308,
|
| 1464 |
+
"eval_steps_per_second": 3.544,
|
| 1465 |
+
"step": 800
|
| 1466 |
+
},
|
| 1467 |
+
{
|
| 1468 |
+
"epoch": 0.8635394456289979,
|
| 1469 |
+
"grad_norm": 1.6861388683319092,
|
| 1470 |
+
"learning_rate": 2.186666916089239e-06,
|
| 1471 |
+
"logits/chosen": 3.2604496479034424,
|
| 1472 |
+
"logits/rejected": 2.558086395263672,
|
| 1473 |
+
"logps/chosen": -3.778765916824341,
|
| 1474 |
+
"logps/rejected": -6.954268455505371,
|
| 1475 |
+
"loss": 0.2923,
|
| 1476 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1477 |
+
"rewards/chosen": -5.668148994445801,
|
| 1478 |
+
"rewards/margins": 4.763254165649414,
|
| 1479 |
+
"rewards/rejected": -10.431403160095215,
|
| 1480 |
+
"step": 810
|
| 1481 |
+
},
|
| 1482 |
+
{
|
| 1483 |
+
"epoch": 0.8742004264392325,
|
| 1484 |
+
"grad_norm": 3.100459575653076,
|
| 1485 |
+
"learning_rate": 2.134792428593971e-06,
|
| 1486 |
+
"logits/chosen": 2.6463425159454346,
|
| 1487 |
+
"logits/rejected": 2.2412638664245605,
|
| 1488 |
+
"logps/chosen": -4.3968634605407715,
|
| 1489 |
+
"logps/rejected": -7.397219181060791,
|
| 1490 |
+
"loss": 0.2767,
|
| 1491 |
+
"rewards/accuracies": 0.862500011920929,
|
| 1492 |
+
"rewards/chosen": -6.595294952392578,
|
| 1493 |
+
"rewards/margins": 4.500533103942871,
|
| 1494 |
+
"rewards/rejected": -11.09582805633545,
|
| 1495 |
+
"step": 820
|
| 1496 |
+
},
|
| 1497 |
+
{
|
| 1498 |
+
"epoch": 0.8848614072494669,
|
| 1499 |
+
"grad_norm": 2.7398672103881836,
|
| 1500 |
+
"learning_rate": 2.0830781332097446e-06,
|
| 1501 |
+
"logits/chosen": 2.8484559059143066,
|
| 1502 |
+
"logits/rejected": 2.408825635910034,
|
| 1503 |
+
"logps/chosen": -3.5274035930633545,
|
| 1504 |
+
"logps/rejected": -6.786914825439453,
|
| 1505 |
+
"loss": 0.2587,
|
| 1506 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 1507 |
+
"rewards/chosen": -5.291105270385742,
|
| 1508 |
+
"rewards/margins": 4.889266014099121,
|
| 1509 |
+
"rewards/rejected": -10.180373191833496,
|
| 1510 |
+
"step": 830
|
| 1511 |
+
},
|
| 1512 |
+
{
|
| 1513 |
+
"epoch": 0.8955223880597015,
|
| 1514 |
+
"grad_norm": 4.221662521362305,
|
| 1515 |
+
"learning_rate": 2.031546713535688e-06,
|
| 1516 |
+
"logits/chosen": 2.5181362628936768,
|
| 1517 |
+
"logits/rejected": 2.1050801277160645,
|
| 1518 |
+
"logps/chosen": -3.7997944355010986,
|
| 1519 |
+
"logps/rejected": -7.4044928550720215,
|
| 1520 |
+
"loss": 0.2985,
|
| 1521 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 1522 |
+
"rewards/chosen": -5.6996917724609375,
|
| 1523 |
+
"rewards/margins": 5.407048225402832,
|
| 1524 |
+
"rewards/rejected": -11.10673999786377,
|
| 1525 |
+
"step": 840
|
| 1526 |
+
},
|
| 1527 |
+
{
|
| 1528 |
+
"epoch": 0.906183368869936,
|
| 1529 |
+
"grad_norm": 4.389741897583008,
|
| 1530 |
+
"learning_rate": 1.9802207729556023e-06,
|
| 1531 |
+
"logits/chosen": 1.7594547271728516,
|
| 1532 |
+
"logits/rejected": 1.5125747919082642,
|
| 1533 |
+
"logps/chosen": -3.7186477184295654,
|
| 1534 |
+
"logps/rejected": -7.241146087646484,
|
| 1535 |
+
"loss": 0.278,
|
| 1536 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1537 |
+
"rewards/chosen": -5.577971935272217,
|
| 1538 |
+
"rewards/margins": 5.28374719619751,
|
| 1539 |
+
"rewards/rejected": -10.861719131469727,
|
| 1540 |
+
"step": 850
|
| 1541 |
+
},
|
| 1542 |
+
{
|
| 1543 |
+
"epoch": 0.906183368869936,
|
| 1544 |
+
"eval_logits/chosen": 2.196138381958008,
|
| 1545 |
+
"eval_logits/rejected": 2.091784715652466,
|
| 1546 |
+
"eval_logps/chosen": -3.610405683517456,
|
| 1547 |
+
"eval_logps/rejected": -7.241917610168457,
|
| 1548 |
+
"eval_loss": 0.2656216025352478,
|
| 1549 |
+
"eval_rewards/accuracies": 0.9473684430122375,
|
| 1550 |
+
"eval_rewards/chosen": -5.415609359741211,
|
| 1551 |
+
"eval_rewards/margins": 5.447267532348633,
|
| 1552 |
+
"eval_rewards/rejected": -10.862876892089844,
|
| 1553 |
+
"eval_runtime": 21.4468,
|
| 1554 |
+
"eval_samples_per_second": 28.303,
|
| 1555 |
+
"eval_steps_per_second": 3.544,
|
| 1556 |
+
"step": 850
|
| 1557 |
+
},
|
| 1558 |
+
{
|
| 1559 |
+
"epoch": 0.9168443496801706,
|
| 1560 |
+
"grad_norm": 4.561154365539551,
|
| 1561 |
+
"learning_rate": 1.9291228247233607e-06,
|
| 1562 |
+
"logits/chosen": 3.5380966663360596,
|
| 1563 |
+
"logits/rejected": 3.1444525718688965,
|
| 1564 |
+
"logps/chosen": -4.1595587730407715,
|
| 1565 |
+
"logps/rejected": -7.205758571624756,
|
| 1566 |
+
"loss": 0.252,
|
| 1567 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 1568 |
+
"rewards/chosen": -6.239338397979736,
|
| 1569 |
+
"rewards/margins": 4.56929874420166,
|
| 1570 |
+
"rewards/rejected": -10.808636665344238,
|
| 1571 |
+
"step": 860
|
| 1572 |
+
},
|
| 1573 |
+
{
|
| 1574 |
+
"epoch": 0.9275053304904051,
|
| 1575 |
+
"grad_norm": 3.8300223350524902,
|
| 1576 |
+
"learning_rate": 1.8782752820878636e-06,
|
| 1577 |
+
"logits/chosen": 2.582847833633423,
|
| 1578 |
+
"logits/rejected": 2.341796398162842,
|
| 1579 |
+
"logps/chosen": -3.9722533226013184,
|
| 1580 |
+
"logps/rejected": -7.717828273773193,
|
| 1581 |
+
"loss": 0.3087,
|
| 1582 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 1583 |
+
"rewards/chosen": -5.958379745483398,
|
| 1584 |
+
"rewards/margins": 5.618363380432129,
|
| 1585 |
+
"rewards/rejected": -11.576744079589844,
|
| 1586 |
+
"step": 870
|
| 1587 |
+
},
|
| 1588 |
+
{
|
| 1589 |
+
"epoch": 0.9381663113006397,
|
| 1590 |
+
"grad_norm": 2.60563063621521,
|
| 1591 |
+
"learning_rate": 1.827700448461836e-06,
|
| 1592 |
+
"logits/chosen": 2.616633892059326,
|
| 1593 |
+
"logits/rejected": 2.4994826316833496,
|
| 1594 |
+
"logps/chosen": -3.9796085357666016,
|
| 1595 |
+
"logps/rejected": -7.014098167419434,
|
| 1596 |
+
"loss": 0.2733,
|
| 1597 |
+
"rewards/accuracies": 0.9375,
|
| 1598 |
+
"rewards/chosen": -5.969412803649902,
|
| 1599 |
+
"rewards/margins": 4.551734924316406,
|
| 1600 |
+
"rewards/rejected": -10.521148681640625,
|
| 1601 |
+
"step": 880
|
| 1602 |
+
},
|
| 1603 |
+
{
|
| 1604 |
+
"epoch": 0.9488272921108742,
|
| 1605 |
+
"grad_norm": 2.454512119293213,
|
| 1606 |
+
"learning_rate": 1.7774205076388207e-06,
|
| 1607 |
+
"logits/chosen": 2.3873534202575684,
|
| 1608 |
+
"logits/rejected": 1.922217607498169,
|
| 1609 |
+
"logps/chosen": -4.740939617156982,
|
| 1610 |
+
"logps/rejected": -8.011380195617676,
|
| 1611 |
+
"loss": 0.2365,
|
| 1612 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1613 |
+
"rewards/chosen": -7.1114091873168945,
|
| 1614 |
+
"rewards/margins": 4.9056596755981445,
|
| 1615 |
+
"rewards/rejected": -12.017068862915039,
|
| 1616 |
+
"step": 890
|
| 1617 |
+
},
|
| 1618 |
+
{
|
| 1619 |
+
"epoch": 0.9594882729211087,
|
| 1620 |
+
"grad_norm": 5.833195209503174,
|
| 1621 |
+
"learning_rate": 1.7274575140626318e-06,
|
| 1622 |
+
"logits/chosen": 2.42203950881958,
|
| 1623 |
+
"logits/rejected": 1.9541574716567993,
|
| 1624 |
+
"logps/chosen": -4.811551094055176,
|
| 1625 |
+
"logps/rejected": -7.7176384925842285,
|
| 1626 |
+
"loss": 0.2397,
|
| 1627 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 1628 |
+
"rewards/chosen": -7.2173261642456055,
|
| 1629 |
+
"rewards/margins": 4.359131813049316,
|
| 1630 |
+
"rewards/rejected": -11.576457977294922,
|
| 1631 |
+
"step": 900
|
| 1632 |
+
},
|
| 1633 |
+
{
|
| 1634 |
+
"epoch": 0.9594882729211087,
|
| 1635 |
+
"eval_logits/chosen": 2.22493052482605,
|
| 1636 |
+
"eval_logits/rejected": 2.157407760620117,
|
| 1637 |
+
"eval_logps/chosen": -3.8421285152435303,
|
| 1638 |
+
"eval_logps/rejected": -7.569239616394043,
|
| 1639 |
+
"eval_loss": 0.2578875720500946,
|
| 1640 |
+
"eval_rewards/accuracies": 0.9342105388641357,
|
| 1641 |
+
"eval_rewards/chosen": -5.763192653656006,
|
| 1642 |
+
"eval_rewards/margins": 5.590666770935059,
|
| 1643 |
+
"eval_rewards/rejected": -11.353858947753906,
|
| 1644 |
+
"eval_runtime": 21.4496,
|
| 1645 |
+
"eval_samples_per_second": 28.299,
|
| 1646 |
+
"eval_steps_per_second": 3.543,
|
| 1647 |
+
"step": 900
|
| 1648 |
+
},
|
| 1649 |
+
{
|
| 1650 |
+
"epoch": 0.9701492537313433,
|
| 1651 |
+
"grad_norm": 2.5518722534179688,
|
| 1652 |
+
"learning_rate": 1.677833383153542e-06,
|
| 1653 |
+
"logits/chosen": 2.4641318321228027,
|
| 1654 |
+
"logits/rejected": 2.093945264816284,
|
| 1655 |
+
"logps/chosen": -4.148250102996826,
|
| 1656 |
+
"logps/rejected": -7.6053924560546875,
|
| 1657 |
+
"loss": 0.2498,
|
| 1658 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1659 |
+
"rewards/chosen": -6.22237491607666,
|
| 1660 |
+
"rewards/margins": 5.185714244842529,
|
| 1661 |
+
"rewards/rejected": -11.408089637756348,
|
| 1662 |
+
"step": 910
|
| 1663 |
+
},
|
| 1664 |
+
{
|
| 1665 |
+
"epoch": 0.9808102345415778,
|
| 1666 |
+
"grad_norm": 3.0121946334838867,
|
| 1667 |
+
"learning_rate": 1.6285698816954626e-06,
|
| 1668 |
+
"logits/chosen": 2.5797393321990967,
|
| 1669 |
+
"logits/rejected": 2.4002528190612793,
|
| 1670 |
+
"logps/chosen": -4.170048236846924,
|
| 1671 |
+
"logps/rejected": -7.2330145835876465,
|
| 1672 |
+
"loss": 0.243,
|
| 1673 |
+
"rewards/accuracies": 0.875,
|
| 1674 |
+
"rewards/chosen": -6.255072116851807,
|
| 1675 |
+
"rewards/margins": 4.594450950622559,
|
| 1676 |
+
"rewards/rejected": -10.849523544311523,
|
| 1677 |
+
"step": 920
|
| 1678 |
+
},
|
| 1679 |
+
{
|
| 1680 |
+
"epoch": 0.9914712153518124,
|
| 1681 |
+
"grad_norm": 6.5740532875061035,
|
| 1682 |
+
"learning_rate": 1.5796886182883053e-06,
|
| 1683 |
+
"logits/chosen": 2.631937265396118,
|
| 1684 |
+
"logits/rejected": 2.3605117797851562,
|
| 1685 |
+
"logps/chosen": -4.277863502502441,
|
| 1686 |
+
"logps/rejected": -7.579499244689941,
|
| 1687 |
+
"loss": 0.2851,
|
| 1688 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1689 |
+
"rewards/chosen": -6.416795253753662,
|
| 1690 |
+
"rewards/margins": 4.952453136444092,
|
| 1691 |
+
"rewards/rejected": -11.369248390197754,
|
| 1692 |
+
"step": 930
|
| 1693 |
+
},
|
| 1694 |
+
{
|
| 1695 |
+
"epoch": 1.0021321961620469,
|
| 1696 |
+
"grad_norm": 3.664747953414917,
|
| 1697 |
+
"learning_rate": 1.5312110338697427e-06,
|
| 1698 |
+
"logits/chosen": 2.602048635482788,
|
| 1699 |
+
"logits/rejected": 2.3019237518310547,
|
| 1700 |
+
"logps/chosen": -4.545152187347412,
|
| 1701 |
+
"logps/rejected": -8.159948348999023,
|
| 1702 |
+
"loss": 0.2524,
|
| 1703 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 1704 |
+
"rewards/chosen": -6.8177289962768555,
|
| 1705 |
+
"rewards/margins": 5.422194004058838,
|
| 1706 |
+
"rewards/rejected": -12.239922523498535,
|
| 1707 |
+
"step": 940
|
| 1708 |
+
},
|
| 1709 |
+
{
|
| 1710 |
+
"epoch": 1.0127931769722816,
|
| 1711 |
+
"grad_norm": 3.0868308544158936,
|
| 1712 |
+
"learning_rate": 1.4831583923105e-06,
|
| 1713 |
+
"logits/chosen": 2.289614677429199,
|
| 1714 |
+
"logits/rejected": 1.9691753387451172,
|
| 1715 |
+
"logps/chosen": -4.014752388000488,
|
| 1716 |
+
"logps/rejected": -7.477097988128662,
|
| 1717 |
+
"loss": 0.2788,
|
| 1718 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1719 |
+
"rewards/chosen": -6.022129058837891,
|
| 1720 |
+
"rewards/margins": 5.19351863861084,
|
| 1721 |
+
"rewards/rejected": -11.215646743774414,
|
| 1722 |
+
"step": 950
|
| 1723 |
+
},
|
| 1724 |
+
{
|
| 1725 |
+
"epoch": 1.0127931769722816,
|
| 1726 |
+
"eval_logits/chosen": 2.137803316116333,
|
| 1727 |
+
"eval_logits/rejected": 2.122602939605713,
|
| 1728 |
+
"eval_logps/chosen": -3.709460496902466,
|
| 1729 |
+
"eval_logps/rejected": -7.5273919105529785,
|
| 1730 |
+
"eval_loss": 0.2563049793243408,
|
| 1731 |
+
"eval_rewards/accuracies": 0.9473684430122375,
|
| 1732 |
+
"eval_rewards/chosen": -5.56419038772583,
|
| 1733 |
+
"eval_rewards/margins": 5.726898670196533,
|
| 1734 |
+
"eval_rewards/rejected": -11.291089057922363,
|
| 1735 |
+
"eval_runtime": 21.4425,
|
| 1736 |
+
"eval_samples_per_second": 28.308,
|
| 1737 |
+
"eval_steps_per_second": 3.544,
|
| 1738 |
+
"step": 950
|
| 1739 |
+
},
|
| 1740 |
+
{
|
| 1741 |
+
"epoch": 1.023454157782516,
|
| 1742 |
+
"grad_norm": 4.523952960968018,
|
| 1743 |
+
"learning_rate": 1.4355517710873184e-06,
|
| 1744 |
+
"logits/chosen": 3.2274162769317627,
|
| 1745 |
+
"logits/rejected": 2.9835076332092285,
|
| 1746 |
+
"logps/chosen": -4.102515697479248,
|
| 1747 |
+
"logps/rejected": -7.5308661460876465,
|
| 1748 |
+
"loss": 0.2394,
|
| 1749 |
+
"rewards/accuracies": 0.887499988079071,
|
| 1750 |
+
"rewards/chosen": -6.153773307800293,
|
| 1751 |
+
"rewards/margins": 5.142527103424072,
|
| 1752 |
+
"rewards/rejected": -11.296300888061523,
|
| 1753 |
+
"step": 960
|
| 1754 |
+
},
|
| 1755 |
+
{
|
| 1756 |
+
"epoch": 1.0341151385927505,
|
| 1757 |
+
"grad_norm": 5.114956378936768,
|
| 1758 |
+
"learning_rate": 1.388412052037682e-06,
|
| 1759 |
+
"logits/chosen": 2.7468860149383545,
|
| 1760 |
+
"logits/rejected": 2.2672038078308105,
|
| 1761 |
+
"logps/chosen": -3.705359935760498,
|
| 1762 |
+
"logps/rejected": -7.7518720626831055,
|
| 1763 |
+
"loss": 0.2419,
|
| 1764 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1765 |
+
"rewards/chosen": -5.558040142059326,
|
| 1766 |
+
"rewards/margins": 6.069769382476807,
|
| 1767 |
+
"rewards/rejected": -11.627809524536133,
|
| 1768 |
+
"step": 970
|
| 1769 |
+
},
|
| 1770 |
+
{
|
| 1771 |
+
"epoch": 1.044776119402985,
|
| 1772 |
+
"grad_norm": 3.243952512741089,
|
| 1773 |
+
"learning_rate": 1.3417599122003464e-06,
|
| 1774 |
+
"logits/chosen": 3.092737913131714,
|
| 1775 |
+
"logits/rejected": 2.6904711723327637,
|
| 1776 |
+
"logps/chosen": -4.205164909362793,
|
| 1777 |
+
"logps/rejected": -7.69910192489624,
|
| 1778 |
+
"loss": 0.275,
|
| 1779 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1780 |
+
"rewards/chosen": -6.307746887207031,
|
| 1781 |
+
"rewards/margins": 5.240906238555908,
|
| 1782 |
+
"rewards/rejected": -11.548653602600098,
|
| 1783 |
+
"step": 980
|
| 1784 |
+
},
|
| 1785 |
+
{
|
| 1786 |
+
"epoch": 1.0554371002132197,
|
| 1787 |
+
"grad_norm": 4.584798812866211,
|
| 1788 |
+
"learning_rate": 1.2956158147457116e-06,
|
| 1789 |
+
"logits/chosen": 1.9939275979995728,
|
| 1790 |
+
"logits/rejected": 1.555945634841919,
|
| 1791 |
+
"logps/chosen": -3.4762721061706543,
|
| 1792 |
+
"logps/rejected": -7.17560338973999,
|
| 1793 |
+
"loss": 0.2523,
|
| 1794 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1795 |
+
"rewards/chosen": -5.2144083976745605,
|
| 1796 |
+
"rewards/margins": 5.548996448516846,
|
| 1797 |
+
"rewards/rejected": -10.763405799865723,
|
| 1798 |
+
"step": 990
|
| 1799 |
+
},
|
| 1800 |
+
{
|
| 1801 |
+
"epoch": 1.0660980810234542,
|
| 1802 |
+
"grad_norm": 3.283021926879883,
|
| 1803 |
+
"learning_rate": 1.2500000000000007e-06,
|
| 1804 |
+
"logits/chosen": 3.2125916481018066,
|
| 1805 |
+
"logits/rejected": 2.6990933418273926,
|
| 1806 |
+
"logps/chosen": -3.818108320236206,
|
| 1807 |
+
"logps/rejected": -7.657105445861816,
|
| 1808 |
+
"loss": 0.281,
|
| 1809 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 1810 |
+
"rewards/chosen": -5.7271623611450195,
|
| 1811 |
+
"rewards/margins": 5.758495330810547,
|
| 1812 |
+
"rewards/rejected": -11.48565673828125,
|
| 1813 |
+
"step": 1000
|
| 1814 |
+
},
|
| 1815 |
+
{
|
| 1816 |
+
"epoch": 1.0660980810234542,
|
| 1817 |
+
"eval_logits/chosen": 2.120778799057007,
|
| 1818 |
+
"eval_logits/rejected": 2.1084940433502197,
|
| 1819 |
+
"eval_logps/chosen": -3.530414342880249,
|
| 1820 |
+
"eval_logps/rejected": -7.398301124572754,
|
| 1821 |
+
"eval_loss": 0.257539838552475,
|
| 1822 |
+
"eval_rewards/accuracies": 0.9473684430122375,
|
| 1823 |
+
"eval_rewards/chosen": -5.295621395111084,
|
| 1824 |
+
"eval_rewards/margins": 5.801830291748047,
|
| 1825 |
+
"eval_rewards/rejected": -11.097451210021973,
|
| 1826 |
+
"eval_runtime": 21.4442,
|
| 1827 |
+
"eval_samples_per_second": 28.306,
|
| 1828 |
+
"eval_steps_per_second": 3.544,
|
| 1829 |
+
"step": 1000
|
| 1830 |
+
},
|
| 1831 |
+
{
|
| 1832 |
+
"epoch": 1.0767590618336886,
|
| 1833 |
+
"grad_norm": 5.394765377044678,
|
| 1834 |
+
"learning_rate": 1.204932476567175e-06,
|
| 1835 |
+
"logits/chosen": 2.473863363265991,
|
| 1836 |
+
"logits/rejected": 2.0679373741149902,
|
| 1837 |
+
"logps/chosen": -4.0380940437316895,
|
| 1838 |
+
"logps/rejected": -7.006314277648926,
|
| 1839 |
+
"loss": 0.2471,
|
| 1840 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 1841 |
+
"rewards/chosen": -6.057140350341797,
|
| 1842 |
+
"rewards/margins": 4.452331066131592,
|
| 1843 |
+
"rewards/rejected": -10.509471893310547,
|
| 1844 |
+
"step": 1010
|
| 1845 |
+
},
|
| 1846 |
+
{
|
| 1847 |
+
"epoch": 1.0874200426439233,
|
| 1848 |
+
"grad_norm": 4.336904525756836,
|
| 1849 |
+
"learning_rate": 1.160433012552508e-06,
|
| 1850 |
+
"logits/chosen": 2.386993408203125,
|
| 1851 |
+
"logits/rejected": 2.060482978820801,
|
| 1852 |
+
"logps/chosen": -3.8012115955352783,
|
| 1853 |
+
"logps/rejected": -7.717658996582031,
|
| 1854 |
+
"loss": 0.2664,
|
| 1855 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1856 |
+
"rewards/chosen": -5.701817512512207,
|
| 1857 |
+
"rewards/margins": 5.874671936035156,
|
| 1858 |
+
"rewards/rejected": -11.576489448547363,
|
| 1859 |
+
"step": 1020
|
| 1860 |
+
},
|
| 1861 |
+
{
|
| 1862 |
+
"epoch": 1.0980810234541578,
|
| 1863 |
+
"grad_norm": 3.5417964458465576,
|
| 1864 |
+
"learning_rate": 1.11652112689164e-06,
|
| 1865 |
+
"logits/chosen": 3.1271402835845947,
|
| 1866 |
+
"logits/rejected": 2.6316893100738525,
|
| 1867 |
+
"logps/chosen": -4.178269386291504,
|
| 1868 |
+
"logps/rejected": -7.592785835266113,
|
| 1869 |
+
"loss": 0.237,
|
| 1870 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1871 |
+
"rewards/chosen": -6.267404556274414,
|
| 1872 |
+
"rewards/margins": 5.121774196624756,
|
| 1873 |
+
"rewards/rejected": -11.389179229736328,
|
| 1874 |
+
"step": 1030
|
| 1875 |
+
},
|
| 1876 |
+
{
|
| 1877 |
+
"epoch": 1.1087420042643923,
|
| 1878 |
+
"grad_norm": 5.173072338104248,
|
| 1879 |
+
"learning_rate": 1.073216080788921e-06,
|
| 1880 |
+
"logits/chosen": 2.7356886863708496,
|
| 1881 |
+
"logits/rejected": 2.5491487979888916,
|
| 1882 |
+
"logps/chosen": -4.352596282958984,
|
| 1883 |
+
"logps/rejected": -7.594882011413574,
|
| 1884 |
+
"loss": 0.2152,
|
| 1885 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 1886 |
+
"rewards/chosen": -6.528894901275635,
|
| 1887 |
+
"rewards/margins": 4.86342716217041,
|
| 1888 |
+
"rewards/rejected": -11.392321586608887,
|
| 1889 |
+
"step": 1040
|
| 1890 |
+
},
|
| 1891 |
+
{
|
| 1892 |
+
"epoch": 1.1194029850746268,
|
| 1893 |
+
"grad_norm": 5.773068904876709,
|
| 1894 |
+
"learning_rate": 1.0305368692688175e-06,
|
| 1895 |
+
"logits/chosen": 2.8095126152038574,
|
| 1896 |
+
"logits/rejected": 2.6465506553649902,
|
| 1897 |
+
"logps/chosen": -4.555963039398193,
|
| 1898 |
+
"logps/rejected": -8.115939140319824,
|
| 1899 |
+
"loss": 0.2337,
|
| 1900 |
+
"rewards/accuracies": 0.887499988079071,
|
| 1901 |
+
"rewards/chosen": -6.833944797515869,
|
| 1902 |
+
"rewards/margins": 5.339962959289551,
|
| 1903 |
+
"rewards/rejected": -12.173909187316895,
|
| 1904 |
+
"step": 1050
|
| 1905 |
+
},
|
| 1906 |
+
{
|
| 1907 |
+
"epoch": 1.1194029850746268,
|
| 1908 |
+
"eval_logits/chosen": 2.186048746109009,
|
| 1909 |
+
"eval_logits/rejected": 2.195524215698242,
|
| 1910 |
+
"eval_logps/chosen": -3.748455286026001,
|
| 1911 |
+
"eval_logps/rejected": -7.740383625030518,
|
| 1912 |
+
"eval_loss": 0.24803213775157928,
|
| 1913 |
+
"eval_rewards/accuracies": 0.9342105388641357,
|
| 1914 |
+
"eval_rewards/chosen": -5.622683525085449,
|
| 1915 |
+
"eval_rewards/margins": 5.987893104553223,
|
| 1916 |
+
"eval_rewards/rejected": -11.610575675964355,
|
| 1917 |
+
"eval_runtime": 21.445,
|
| 1918 |
+
"eval_samples_per_second": 28.305,
|
| 1919 |
+
"eval_steps_per_second": 3.544,
|
| 1920 |
+
"step": 1050
|
| 1921 |
+
},
|
| 1922 |
+
{
|
| 1923 |
+
"epoch": 1.1300639658848615,
|
| 1924 |
+
"grad_norm": 3.7594547271728516,
|
| 1925 |
+
"learning_rate": 9.88502212844063e-07,
|
| 1926 |
+
"logits/chosen": 2.6019225120544434,
|
| 1927 |
+
"logits/rejected": 2.0351009368896484,
|
| 1928 |
+
"logps/chosen": -3.759199857711792,
|
| 1929 |
+
"logps/rejected": -7.5345635414123535,
|
| 1930 |
+
"loss": 0.226,
|
| 1931 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 1932 |
+
"rewards/chosen": -5.638800621032715,
|
| 1933 |
+
"rewards/margins": 5.6630449295043945,
|
| 1934 |
+
"rewards/rejected": -11.301844596862793,
|
| 1935 |
+
"step": 1060
|
| 1936 |
+
},
|
| 1937 |
+
{
|
| 1938 |
+
"epoch": 1.140724946695096,
|
| 1939 |
+
"grad_norm": 4.682251453399658,
|
| 1940 |
+
"learning_rate": 9.471305493042243e-07,
|
| 1941 |
+
"logits/chosen": 2.9025776386260986,
|
| 1942 |
+
"logits/rejected": 2.5837249755859375,
|
| 1943 |
+
"logps/chosen": -4.261716842651367,
|
| 1944 |
+
"logps/rejected": -8.397085189819336,
|
| 1945 |
+
"loss": 0.2819,
|
| 1946 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 1947 |
+
"rewards/chosen": -6.392575263977051,
|
| 1948 |
+
"rewards/margins": 6.2030534744262695,
|
| 1949 |
+
"rewards/rejected": -12.59562873840332,
|
| 1950 |
+
"step": 1070
|
| 1951 |
+
},
|
| 1952 |
+
{
|
| 1953 |
+
"epoch": 1.1513859275053304,
|
| 1954 |
+
"grad_norm": 4.629392147064209,
|
| 1955 |
+
"learning_rate": 9.064400256282757e-07,
|
| 1956 |
+
"logits/chosen": 2.7138123512268066,
|
| 1957 |
+
"logits/rejected": 2.5913774967193604,
|
| 1958 |
+
"logps/chosen": -4.063493251800537,
|
| 1959 |
+
"logps/rejected": -7.581247806549072,
|
| 1960 |
+
"loss": 0.2449,
|
| 1961 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 1962 |
+
"rewards/chosen": -6.095240116119385,
|
| 1963 |
+
"rewards/margins": 5.2766313552856445,
|
| 1964 |
+
"rewards/rejected": -11.371870994567871,
|
| 1965 |
+
"step": 1080
|
| 1966 |
+
},
|
| 1967 |
+
{
|
| 1968 |
+
"epoch": 1.1620469083155651,
|
| 1969 |
+
"grad_norm": 3.096658706665039,
|
| 1970 |
+
"learning_rate": 8.664484900247363e-07,
|
| 1971 |
+
"logits/chosen": 2.294330596923828,
|
| 1972 |
+
"logits/rejected": 1.9075424671173096,
|
| 1973 |
+
"logps/chosen": -4.294546604156494,
|
| 1974 |
+
"logps/rejected": -8.126740455627441,
|
| 1975 |
+
"loss": 0.2137,
|
| 1976 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 1977 |
+
"rewards/chosen": -6.441819667816162,
|
| 1978 |
+
"rewards/margins": 5.748291492462158,
|
| 1979 |
+
"rewards/rejected": -12.19011116027832,
|
| 1980 |
+
"step": 1090
|
| 1981 |
+
},
|
| 1982 |
+
{
|
| 1983 |
+
"epoch": 1.1727078891257996,
|
| 1984 |
+
"grad_norm": 7.750303268432617,
|
| 1985 |
+
"learning_rate": 8.271734841028553e-07,
|
| 1986 |
+
"logits/chosen": 2.7251312732696533,
|
| 1987 |
+
"logits/rejected": 1.9645192623138428,
|
| 1988 |
+
"logps/chosen": -4.135927677154541,
|
| 1989 |
+
"logps/rejected": -7.487878322601318,
|
| 1990 |
+
"loss": 0.2413,
|
| 1991 |
+
"rewards/accuracies": 0.9375,
|
| 1992 |
+
"rewards/chosen": -6.203891277313232,
|
| 1993 |
+
"rewards/margins": 5.027926445007324,
|
| 1994 |
+
"rewards/rejected": -11.231817245483398,
|
| 1995 |
+
"step": 1100
|
| 1996 |
+
},
|
| 1997 |
+
{
|
| 1998 |
+
"epoch": 1.1727078891257996,
|
| 1999 |
+
"eval_logits/chosen": 2.114663600921631,
|
| 2000 |
+
"eval_logits/rejected": 2.1521003246307373,
|
| 2001 |
+
"eval_logps/chosen": -3.7643239498138428,
|
| 2002 |
+
"eval_logps/rejected": -7.814781665802002,
|
| 2003 |
+
"eval_loss": 0.24271255731582642,
|
| 2004 |
+
"eval_rewards/accuracies": 0.9342105388641357,
|
| 2005 |
+
"eval_rewards/chosen": -5.646485805511475,
|
| 2006 |
+
"eval_rewards/margins": 6.075687408447266,
|
| 2007 |
+
"eval_rewards/rejected": -11.722171783447266,
|
| 2008 |
+
"eval_runtime": 21.4461,
|
| 2009 |
+
"eval_samples_per_second": 28.304,
|
| 2010 |
+
"eval_steps_per_second": 3.544,
|
| 2011 |
+
"step": 1100
|
| 2012 |
+
},
|
| 2013 |
+
{
|
| 2014 |
+
"epoch": 1.183368869936034,
|
| 2015 |
+
"grad_norm": 3.1332218647003174,
|
| 2016 |
+
"learning_rate": 7.886322351782782e-07,
|
| 2017 |
+
"logits/chosen": 2.4613986015319824,
|
| 2018 |
+
"logits/rejected": 2.4474260807037354,
|
| 2019 |
+
"logps/chosen": -4.631771564483643,
|
| 2020 |
+
"logps/rejected": -8.386178970336914,
|
| 2021 |
+
"loss": 0.2099,
|
| 2022 |
+
"rewards/accuracies": 0.9375,
|
| 2023 |
+
"rewards/chosen": -6.947657108306885,
|
| 2024 |
+
"rewards/margins": 5.631610870361328,
|
| 2025 |
+
"rewards/rejected": -12.579267501831055,
|
| 2026 |
+
"step": 1110
|
| 2027 |
+
},
|
| 2028 |
+
{
|
| 2029 |
+
"epoch": 1.1940298507462686,
|
| 2030 |
+
"grad_norm": 3.8220393657684326,
|
| 2031 |
+
"learning_rate": 7.508416487165862e-07,
|
| 2032 |
+
"logits/chosen": 1.7835853099822998,
|
| 2033 |
+
"logits/rejected": 1.3680188655853271,
|
| 2034 |
+
"logps/chosen": -3.990460157394409,
|
| 2035 |
+
"logps/rejected": -7.8566789627075195,
|
| 2036 |
+
"loss": 0.2527,
|
| 2037 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 2038 |
+
"rewards/chosen": -5.985690116882324,
|
| 2039 |
+
"rewards/margins": 5.799327850341797,
|
| 2040 |
+
"rewards/rejected": -11.785018920898438,
|
| 2041 |
+
"step": 1120
|
| 2042 |
+
},
|
| 2043 |
+
{
|
| 2044 |
+
"epoch": 1.2046908315565032,
|
| 2045 |
+
"grad_norm": 4.16413688659668,
|
| 2046 |
+
"learning_rate": 7.138183009179922e-07,
|
| 2047 |
+
"logits/chosen": 2.3430938720703125,
|
| 2048 |
+
"logits/rejected": 1.9356731176376343,
|
| 2049 |
+
"logps/chosen": -3.891129970550537,
|
| 2050 |
+
"logps/rejected": -7.389913082122803,
|
| 2051 |
+
"loss": 0.2229,
|
| 2052 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 2053 |
+
"rewards/chosen": -5.836695194244385,
|
| 2054 |
+
"rewards/margins": 5.248173713684082,
|
| 2055 |
+
"rewards/rejected": -11.084868431091309,
|
| 2056 |
+
"step": 1130
|
| 2057 |
+
},
|
| 2058 |
+
{
|
| 2059 |
+
"epoch": 1.2153518123667377,
|
| 2060 |
+
"grad_norm": 5.904109954833984,
|
| 2061 |
+
"learning_rate": 6.775784314464717e-07,
|
| 2062 |
+
"logits/chosen": 2.3378713130950928,
|
| 2063 |
+
"logits/rejected": 2.5543103218078613,
|
| 2064 |
+
"logps/chosen": -4.343329906463623,
|
| 2065 |
+
"logps/rejected": -7.562542915344238,
|
| 2066 |
+
"loss": 0.2684,
|
| 2067 |
+
"rewards/accuracies": 0.887499988079071,
|
| 2068 |
+
"rewards/chosen": -6.5149946212768555,
|
| 2069 |
+
"rewards/margins": 4.828819751739502,
|
| 2070 |
+
"rewards/rejected": -11.3438138961792,
|
| 2071 |
+
"step": 1140
|
| 2072 |
+
},
|
| 2073 |
+
{
|
| 2074 |
+
"epoch": 1.2260127931769722,
|
| 2075 |
+
"grad_norm": 3.845862865447998,
|
| 2076 |
+
"learning_rate": 6.421379363065142e-07,
|
| 2077 |
+
"logits/chosen": 3.033965587615967,
|
| 2078 |
+
"logits/rejected": 2.835658550262451,
|
| 2079 |
+
"logps/chosen": -4.35538911819458,
|
| 2080 |
+
"logps/rejected": -8.115766525268555,
|
| 2081 |
+
"loss": 0.2625,
|
| 2082 |
+
"rewards/accuracies": 0.949999988079071,
|
| 2083 |
+
"rewards/chosen": -6.533083438873291,
|
| 2084 |
+
"rewards/margins": 5.640565395355225,
|
| 2085 |
+
"rewards/rejected": -12.173648834228516,
|
| 2086 |
+
"step": 1150
|
| 2087 |
+
},
|
| 2088 |
+
{
|
| 2089 |
+
"epoch": 1.2260127931769722,
|
| 2090 |
+
"eval_logits/chosen": 2.2132980823516846,
|
| 2091 |
+
"eval_logits/rejected": 2.2383644580841064,
|
| 2092 |
+
"eval_logps/chosen": -3.8321533203125,
|
| 2093 |
+
"eval_logps/rejected": -7.898870944976807,
|
| 2094 |
+
"eval_loss": 0.23817622661590576,
|
| 2095 |
+
"eval_rewards/accuracies": 0.9342105388641357,
|
| 2096 |
+
"eval_rewards/chosen": -5.74822998046875,
|
| 2097 |
+
"eval_rewards/margins": 6.100078105926514,
|
| 2098 |
+
"eval_rewards/rejected": -11.848306655883789,
|
| 2099 |
+
"eval_runtime": 21.4432,
|
| 2100 |
+
"eval_samples_per_second": 28.307,
|
| 2101 |
+
"eval_steps_per_second": 3.544,
|
| 2102 |
+
"step": 1150
|
| 2103 |
+
},
|
| 2104 |
+
{
|
| 2105 |
+
"epoch": 1.236673773987207,
|
| 2106 |
+
"grad_norm": 3.4539601802825928,
|
| 2107 |
+
"learning_rate": 6.075123608706093e-07,
|
| 2108 |
+
"logits/chosen": 2.93648099899292,
|
| 2109 |
+
"logits/rejected": 2.8411507606506348,
|
| 2110 |
+
"logps/chosen": -4.176449775695801,
|
| 2111 |
+
"logps/rejected": -7.605371952056885,
|
| 2112 |
+
"loss": 0.2479,
|
| 2113 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 2114 |
+
"rewards/chosen": -6.264674186706543,
|
| 2115 |
+
"rewards/margins": 5.143383979797363,
|
| 2116 |
+
"rewards/rejected": -11.408058166503906,
|
| 2117 |
+
"step": 1160
|
| 2118 |
+
},
|
| 2119 |
+
{
|
| 2120 |
+
"epoch": 1.2473347547974414,
|
| 2121 |
+
"grad_norm": 3.6077513694763184,
|
| 2122 |
+
"learning_rate": 5.737168930605272e-07,
|
| 2123 |
+
"logits/chosen": 3.048224925994873,
|
| 2124 |
+
"logits/rejected": 3.061757802963257,
|
| 2125 |
+
"logps/chosen": -4.667507171630859,
|
| 2126 |
+
"logps/rejected": -8.997530937194824,
|
| 2127 |
+
"loss": 0.2096,
|
| 2128 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 2129 |
+
"rewards/chosen": -7.001260280609131,
|
| 2130 |
+
"rewards/margins": 6.4950361251831055,
|
| 2131 |
+
"rewards/rejected": -13.496296882629395,
|
| 2132 |
+
"step": 1170
|
| 2133 |
+
},
|
| 2134 |
+
{
|
| 2135 |
+
"epoch": 1.2579957356076759,
|
| 2136 |
+
"grad_norm": 3.400186061859131,
|
| 2137 |
+
"learning_rate": 5.407663566854008e-07,
|
| 2138 |
+
"logits/chosen": 2.4676003456115723,
|
| 2139 |
+
"logits/rejected": 2.2479636669158936,
|
| 2140 |
+
"logps/chosen": -3.8468101024627686,
|
| 2141 |
+
"logps/rejected": -8.152502059936523,
|
| 2142 |
+
"loss": 0.2542,
|
| 2143 |
+
"rewards/accuracies": 0.925000011920929,
|
| 2144 |
+
"rewards/chosen": -5.7702155113220215,
|
| 2145 |
+
"rewards/margins": 6.458538055419922,
|
| 2146 |
+
"rewards/rejected": -12.228754043579102,
|
| 2147 |
+
"step": 1180
|
| 2148 |
+
},
|
| 2149 |
+
{
|
| 2150 |
+
"epoch": 1.2686567164179103,
|
| 2151 |
+
"grad_norm": 4.976436614990234,
|
| 2152 |
+
"learning_rate": 5.086752049395094e-07,
|
| 2153 |
+
"logits/chosen": 2.905099868774414,
|
| 2154 |
+
"logits/rejected": 2.561856746673584,
|
| 2155 |
+
"logps/chosen": -4.313426971435547,
|
| 2156 |
+
"logps/rejected": -8.205011367797852,
|
| 2157 |
+
"loss": 0.26,
|
| 2158 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 2159 |
+
"rewards/chosen": -6.4701409339904785,
|
| 2160 |
+
"rewards/margins": 5.837375640869141,
|
| 2161 |
+
"rewards/rejected": -12.307516098022461,
|
| 2162 |
+
"step": 1190
|
| 2163 |
+
},
|
| 2164 |
+
{
|
| 2165 |
+
"epoch": 1.279317697228145,
|
| 2166 |
+
"grad_norm": 3.4499430656433105,
|
| 2167 |
+
"learning_rate": 4.774575140626317e-07,
|
| 2168 |
+
"logits/chosen": 2.9027159214019775,
|
| 2169 |
+
"logits/rejected": 2.7010762691497803,
|
| 2170 |
+
"logps/chosen": -4.515786170959473,
|
| 2171 |
+
"logps/rejected": -7.926545143127441,
|
| 2172 |
+
"loss": 0.2364,
|
| 2173 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 2174 |
+
"rewards/chosen": -6.773678779602051,
|
| 2175 |
+
"rewards/margins": 5.116138935089111,
|
| 2176 |
+
"rewards/rejected": -11.88981819152832,
|
| 2177 |
+
"step": 1200
|
| 2178 |
+
},
|
| 2179 |
+
{
|
| 2180 |
+
"epoch": 1.279317697228145,
|
| 2181 |
+
"eval_logits/chosen": 2.23946213722229,
|
| 2182 |
+
"eval_logits/rejected": 2.2642405033111572,
|
| 2183 |
+
"eval_logps/chosen": -3.853114128112793,
|
| 2184 |
+
"eval_logps/rejected": -7.962078094482422,
|
| 2185 |
+
"eval_loss": 0.23450209200382233,
|
| 2186 |
+
"eval_rewards/accuracies": 0.9342105388641357,
|
| 2187 |
+
"eval_rewards/chosen": -5.7796711921691895,
|
| 2188 |
+
"eval_rewards/margins": 6.163444519042969,
|
| 2189 |
+
"eval_rewards/rejected": -11.943115234375,
|
| 2190 |
+
"eval_runtime": 21.4408,
|
| 2191 |
+
"eval_samples_per_second": 28.311,
|
| 2192 |
+
"eval_steps_per_second": 3.545,
|
| 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": 2.8773976633245696e+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 sha256:29030fd2c9579a53d4cbb5166f67721176cc7b02c89ae9a2d5722caa76c57c96
|
| 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)
|