Training in progress, step 1400, checkpoint
Browse files- checkpoint-1400/README.md +202 -0
- checkpoint-1400/adapter_config.json +34 -0
- checkpoint-1400/adapter_model.safetensors +3 -0
- checkpoint-1400/global_step1398/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt +3 -0
- checkpoint-1400/global_step1398/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt +3 -0
- checkpoint-1400/global_step1398/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt +3 -0
- checkpoint-1400/global_step1398/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt +3 -0
- checkpoint-1400/global_step1398/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt +3 -0
- checkpoint-1400/global_step1398/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt +3 -0
- checkpoint-1400/global_step1398/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt +3 -0
- checkpoint-1400/global_step1398/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt +3 -0
- checkpoint-1400/global_step1398/mp_rank_00_model_states.pt +3 -0
- checkpoint-1400/latest +1 -0
- checkpoint-1400/rng_state_0.pth +3 -0
- checkpoint-1400/rng_state_1.pth +3 -0
- checkpoint-1400/rng_state_2.pth +3 -0
- checkpoint-1400/rng_state_3.pth +3 -0
- checkpoint-1400/rng_state_4.pth +3 -0
- checkpoint-1400/rng_state_5.pth +3 -0
- checkpoint-1400/rng_state_6.pth +3 -0
- checkpoint-1400/rng_state_7.pth +3 -0
- checkpoint-1400/scheduler.pt +3 -0
- checkpoint-1400/special_tokens_map.json +30 -0
- checkpoint-1400/tokenizer.json +0 -0
- checkpoint-1400/tokenizer_config.json +133 -0
- checkpoint-1400/trainer_state.json +2581 -0
- checkpoint-1400/training_args.bin +3 -0
- checkpoint-1400/zero_to_fp32.py +674 -0
checkpoint-1400/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-1400/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 |
+
"down_proj",
|
| 27 |
+
"qkv_proj",
|
| 28 |
+
"o_proj",
|
| 29 |
+
"gate_up_proj"
|
| 30 |
+
],
|
| 31 |
+
"task_type": "CAUSAL_LM",
|
| 32 |
+
"use_dora": false,
|
| 33 |
+
"use_rslora": false
|
| 34 |
+
}
|
checkpoint-1400/adapter_model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d06756c8d92882550024f6c457438a61dc700ef3a3d9e0a489eb56a6d771e3fc
|
| 3 |
+
size 25200088
|
checkpoint-1400/global_step1398/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:dfae0d1e65f53bbaa7857718021b41d9b980f6beff73a1ed57c481e436a0c822
|
| 3 |
+
size 18881328
|
checkpoint-1400/global_step1398/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:dc15e7e159099bddc79bb490fd09251aec95760b1c70ce57cf047f6e3aa846bf
|
| 3 |
+
size 18881328
|
checkpoint-1400/global_step1398/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:e0904b93ee6d8434a75c2209c964898fec0c64f6a10ca9b78a2ce4c59f18468f
|
| 3 |
+
size 18881328
|
checkpoint-1400/global_step1398/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 sha256:2226e3584a537766296c3d64c1c557c40a3832fc26ad7e9254b469c69f83de85
|
| 3 |
+
size 18881392
|
checkpoint-1400/global_step1398/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 sha256:17906c9775a85fc5f6c95e42971f4f22326f454a5dcfc99d1eab9769730d8c16
|
| 3 |
+
size 18881392
|
checkpoint-1400/global_step1398/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:e9cf919a4dc17cbc9ec84da76800efc2584806310be9ea500df04161f13f95fb
|
| 3 |
+
size 18881392
|
checkpoint-1400/global_step1398/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 sha256:2bdfc4c2195e16eafaa084d9c369d634835787ced6615e7eacafb1f7bc8fc14b
|
| 3 |
+
size 18881392
|
checkpoint-1400/global_step1398/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 sha256:8632840769126219e8adc03ed0b9de1160d9529cb947999cc06d784b76126828
|
| 3 |
+
size 18881392
|
checkpoint-1400/global_step1398/mp_rank_00_model_states.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:fa748062c20e6302261b37306b33fee8ef73906106245543e270aa402ef3908c
|
| 3 |
+
size 25379244
|
checkpoint-1400/latest
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
global_step1398
|
checkpoint-1400/rng_state_0.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:27670a2e80ef6eea2112fb593ee34bfc11ac56deececdc41375268836746484a
|
| 3 |
+
size 15984
|
checkpoint-1400/rng_state_1.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a9fb368b8e1a011c7ffe983f562eba811ba532df543052875a13820074278b90
|
| 3 |
+
size 15984
|
checkpoint-1400/rng_state_2.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1861c92b516c0a288138c867ba18f97edb9adb30cd5750b68b932248cf8d5557
|
| 3 |
+
size 15984
|
checkpoint-1400/rng_state_3.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e741f8420186fd289dd05fd1db88f41e0da10203461344518d3b8bc30d50b96b
|
| 3 |
+
size 15984
|
checkpoint-1400/rng_state_4.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:88efc65eaf65645838644561e78d57941a96e07300aa390bd886817b7ff8bea7
|
| 3 |
+
size 15984
|
checkpoint-1400/rng_state_5.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8028ddfb7d72b0c282e45ca3a0cadcd5d0bd77db0bb6f3e3c6102423e14b55ae
|
| 3 |
+
size 15984
|
checkpoint-1400/rng_state_6.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8a77febdde7a01f90b20ef76acc76714f772da6245731bf288cd9ea69e57c5fd
|
| 3 |
+
size 15984
|
checkpoint-1400/rng_state_7.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d676e14f0aa43cf9b2bd7f5d1a09749f6dc54c870f621fe9fe6d8e3f08c55169
|
| 3 |
+
size 15984
|
checkpoint-1400/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:46ccd2eaabf2e62001ccbd9f6336cb2a4a520c54a3ad99c9dbcf64407d925ac8
|
| 3 |
+
size 1064
|
checkpoint-1400/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-1400/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
checkpoint-1400/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-1400/trainer_state.json
ADDED
|
@@ -0,0 +1,2581 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 2.194624288797332,
|
| 5 |
+
"eval_steps": 50,
|
| 6 |
+
"global_step": 1400,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.015695507161075144,
|
| 13 |
+
"grad_norm": 0.04355761408805847,
|
| 14 |
+
"learning_rate": 4.999451708687114e-06,
|
| 15 |
+
"logits/chosen": 14.845781326293945,
|
| 16 |
+
"logits/rejected": 14.576438903808594,
|
| 17 |
+
"logps/chosen": -0.31864267587661743,
|
| 18 |
+
"logps/rejected": -0.24545662105083466,
|
| 19 |
+
"loss": 1.0492,
|
| 20 |
+
"rewards/accuracies": 0.25,
|
| 21 |
+
"rewards/chosen": -0.47796401381492615,
|
| 22 |
+
"rewards/margins": -0.10977902263402939,
|
| 23 |
+
"rewards/rejected": -0.3681849539279938,
|
| 24 |
+
"step": 10
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.03139101432215029,
|
| 28 |
+
"grad_norm": 0.04919258877635002,
|
| 29 |
+
"learning_rate": 4.997807075247147e-06,
|
| 30 |
+
"logits/chosen": 15.27595043182373,
|
| 31 |
+
"logits/rejected": 14.872761726379395,
|
| 32 |
+
"logps/chosen": -0.3344747722148895,
|
| 33 |
+
"logps/rejected": -0.24258682131767273,
|
| 34 |
+
"loss": 1.0487,
|
| 35 |
+
"rewards/accuracies": 0.16249999403953552,
|
| 36 |
+
"rewards/chosen": -0.5017121434211731,
|
| 37 |
+
"rewards/margins": -0.1378319263458252,
|
| 38 |
+
"rewards/rejected": -0.3638802468776703,
|
| 39 |
+
"step": 20
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.047086521483225424,
|
| 43 |
+
"grad_norm": 0.049933061003685,
|
| 44 |
+
"learning_rate": 4.9950668210706795e-06,
|
| 45 |
+
"logits/chosen": 15.913165092468262,
|
| 46 |
+
"logits/rejected": 15.607622146606445,
|
| 47 |
+
"logps/chosen": -0.3440183997154236,
|
| 48 |
+
"logps/rejected": -0.2831566333770752,
|
| 49 |
+
"loss": 1.0405,
|
| 50 |
+
"rewards/accuracies": 0.1875,
|
| 51 |
+
"rewards/chosen": -0.516027569770813,
|
| 52 |
+
"rewards/margins": -0.09129264950752258,
|
| 53 |
+
"rewards/rejected": -0.4247349202632904,
|
| 54 |
+
"step": 30
|
| 55 |
+
},
|
| 56 |
+
{
|
| 57 |
+
"epoch": 0.06278202864430057,
|
| 58 |
+
"grad_norm": 0.05693503096699715,
|
| 59 |
+
"learning_rate": 4.9912321481237616e-06,
|
| 60 |
+
"logits/chosen": 15.402900695800781,
|
| 61 |
+
"logits/rejected": 14.99272632598877,
|
| 62 |
+
"logps/chosen": -0.3297731578350067,
|
| 63 |
+
"logps/rejected": -0.2746916711330414,
|
| 64 |
+
"loss": 1.0369,
|
| 65 |
+
"rewards/accuracies": 0.21250000596046448,
|
| 66 |
+
"rewards/chosen": -0.49465981125831604,
|
| 67 |
+
"rewards/margins": -0.08262218534946442,
|
| 68 |
+
"rewards/rejected": -0.41203755140304565,
|
| 69 |
+
"step": 40
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 0.07847753580537571,
|
| 73 |
+
"grad_norm": 0.05467928573489189,
|
| 74 |
+
"learning_rate": 4.986304738420684e-06,
|
| 75 |
+
"logits/chosen": 15.64543342590332,
|
| 76 |
+
"logits/rejected": 15.632547378540039,
|
| 77 |
+
"logps/chosen": -0.30952686071395874,
|
| 78 |
+
"logps/rejected": -0.24847058951854706,
|
| 79 |
+
"loss": 1.0367,
|
| 80 |
+
"rewards/accuracies": 0.21250000596046448,
|
| 81 |
+
"rewards/chosen": -0.46429023146629333,
|
| 82 |
+
"rewards/margins": -0.09158438444137573,
|
| 83 |
+
"rewards/rejected": -0.37270587682724,
|
| 84 |
+
"step": 50
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 0.07847753580537571,
|
| 88 |
+
"eval_logits/chosen": 15.850138664245605,
|
| 89 |
+
"eval_logits/rejected": 15.368529319763184,
|
| 90 |
+
"eval_logps/chosen": -0.3222965598106384,
|
| 91 |
+
"eval_logps/rejected": -0.26877468824386597,
|
| 92 |
+
"eval_loss": 1.0326261520385742,
|
| 93 |
+
"eval_rewards/accuracies": 0.26923078298568726,
|
| 94 |
+
"eval_rewards/chosen": -0.4834447503089905,
|
| 95 |
+
"eval_rewards/margins": -0.08028276264667511,
|
| 96 |
+
"eval_rewards/rejected": -0.40316200256347656,
|
| 97 |
+
"eval_runtime": 14.5044,
|
| 98 |
+
"eval_samples_per_second": 28.405,
|
| 99 |
+
"eval_steps_per_second": 3.585,
|
| 100 |
+
"step": 50
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.09417304296645085,
|
| 104 |
+
"grad_norm": 0.06174452602863312,
|
| 105 |
+
"learning_rate": 4.980286753286196e-06,
|
| 106 |
+
"logits/chosen": 15.443066596984863,
|
| 107 |
+
"logits/rejected": 15.192205429077148,
|
| 108 |
+
"logps/chosen": -0.31090402603149414,
|
| 109 |
+
"logps/rejected": -0.26281923055648804,
|
| 110 |
+
"loss": 1.04,
|
| 111 |
+
"rewards/accuracies": 0.25,
|
| 112 |
+
"rewards/chosen": -0.466356098651886,
|
| 113 |
+
"rewards/margins": -0.07212716341018677,
|
| 114 |
+
"rewards/rejected": -0.39422887563705444,
|
| 115 |
+
"step": 60
|
| 116 |
+
},
|
| 117 |
+
{
|
| 118 |
+
"epoch": 0.109868550127526,
|
| 119 |
+
"grad_norm": 0.06952528655529022,
|
| 120 |
+
"learning_rate": 4.973180832407471e-06,
|
| 121 |
+
"logits/chosen": 16.024200439453125,
|
| 122 |
+
"logits/rejected": 15.82934284210205,
|
| 123 |
+
"logps/chosen": -0.348227322101593,
|
| 124 |
+
"logps/rejected": -0.26220566034317017,
|
| 125 |
+
"loss": 1.043,
|
| 126 |
+
"rewards/accuracies": 0.21250000596046448,
|
| 127 |
+
"rewards/chosen": -0.5223408937454224,
|
| 128 |
+
"rewards/margins": -0.1290324479341507,
|
| 129 |
+
"rewards/rejected": -0.39330852031707764,
|
| 130 |
+
"step": 70
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 0.12556405728860115,
|
| 134 |
+
"grad_norm": 0.07572082430124283,
|
| 135 |
+
"learning_rate": 4.964990092676263e-06,
|
| 136 |
+
"logits/chosen": 15.884915351867676,
|
| 137 |
+
"logits/rejected": 15.603845596313477,
|
| 138 |
+
"logps/chosen": -0.34849274158477783,
|
| 139 |
+
"logps/rejected": -0.26585355401039124,
|
| 140 |
+
"loss": 1.0285,
|
| 141 |
+
"rewards/accuracies": 0.1875,
|
| 142 |
+
"rewards/chosen": -0.5227391719818115,
|
| 143 |
+
"rewards/margins": -0.12395882606506348,
|
| 144 |
+
"rewards/rejected": -0.39878037571907043,
|
| 145 |
+
"step": 80
|
| 146 |
+
},
|
| 147 |
+
{
|
| 148 |
+
"epoch": 0.14125956444967627,
|
| 149 |
+
"grad_norm": 0.2423778474330902,
|
| 150 |
+
"learning_rate": 4.9557181268217225e-06,
|
| 151 |
+
"logits/chosen": 15.978216171264648,
|
| 152 |
+
"logits/rejected": 15.76471996307373,
|
| 153 |
+
"logps/chosen": -0.327436238527298,
|
| 154 |
+
"logps/rejected": -0.25457051396369934,
|
| 155 |
+
"loss": 1.03,
|
| 156 |
+
"rewards/accuracies": 0.22499999403953552,
|
| 157 |
+
"rewards/chosen": -0.49115434288978577,
|
| 158 |
+
"rewards/margins": -0.10929858684539795,
|
| 159 |
+
"rewards/rejected": -0.3818557560443878,
|
| 160 |
+
"step": 90
|
| 161 |
+
},
|
| 162 |
+
{
|
| 163 |
+
"epoch": 0.15695507161075142,
|
| 164 |
+
"grad_norm": 0.1594536453485489,
|
| 165 |
+
"learning_rate": 4.9453690018345144e-06,
|
| 166 |
+
"logits/chosen": 16.307537078857422,
|
| 167 |
+
"logits/rejected": 16.138330459594727,
|
| 168 |
+
"logps/chosen": -0.3342314660549164,
|
| 169 |
+
"logps/rejected": -0.27582648396492004,
|
| 170 |
+
"loss": 1.0309,
|
| 171 |
+
"rewards/accuracies": 0.26249998807907104,
|
| 172 |
+
"rewards/chosen": -0.5013472437858582,
|
| 173 |
+
"rewards/margins": -0.0876075029373169,
|
| 174 |
+
"rewards/rejected": -0.41373974084854126,
|
| 175 |
+
"step": 100
|
| 176 |
+
},
|
| 177 |
+
{
|
| 178 |
+
"epoch": 0.15695507161075142,
|
| 179 |
+
"eval_logits/chosen": 16.4310245513916,
|
| 180 |
+
"eval_logits/rejected": 15.98912525177002,
|
| 181 |
+
"eval_logps/chosen": -0.3239763677120209,
|
| 182 |
+
"eval_logps/rejected": -0.28784558176994324,
|
| 183 |
+
"eval_loss": 1.020836353302002,
|
| 184 |
+
"eval_rewards/accuracies": 0.3076923191547394,
|
| 185 |
+
"eval_rewards/chosen": -0.4859645664691925,
|
| 186 |
+
"eval_rewards/margins": -0.054196178913116455,
|
| 187 |
+
"eval_rewards/rejected": -0.43176835775375366,
|
| 188 |
+
"eval_runtime": 14.5049,
|
| 189 |
+
"eval_samples_per_second": 28.404,
|
| 190 |
+
"eval_steps_per_second": 3.585,
|
| 191 |
+
"step": 100
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.17265057877182657,
|
| 195 |
+
"grad_norm": 0.07431349903345108,
|
| 196 |
+
"learning_rate": 4.933947257182901e-06,
|
| 197 |
+
"logits/chosen": 16.56686782836914,
|
| 198 |
+
"logits/rejected": 16.093189239501953,
|
| 199 |
+
"logps/chosen": -0.34455060958862305,
|
| 200 |
+
"logps/rejected": -0.2834388315677643,
|
| 201 |
+
"loss": 1.0388,
|
| 202 |
+
"rewards/accuracies": 0.30000001192092896,
|
| 203 |
+
"rewards/chosen": -0.5168259739875793,
|
| 204 |
+
"rewards/margins": -0.09166768193244934,
|
| 205 |
+
"rewards/rejected": -0.4251582622528076,
|
| 206 |
+
"step": 110
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.1883460859329017,
|
| 210 |
+
"grad_norm": 0.08802352845668793,
|
| 211 |
+
"learning_rate": 4.921457902821578e-06,
|
| 212 |
+
"logits/chosen": 16.50200843811035,
|
| 213 |
+
"logits/rejected": 16.286388397216797,
|
| 214 |
+
"logps/chosen": -0.30845317244529724,
|
| 215 |
+
"logps/rejected": -0.2677682936191559,
|
| 216 |
+
"loss": 1.0247,
|
| 217 |
+
"rewards/accuracies": 0.25,
|
| 218 |
+
"rewards/chosen": -0.4626797139644623,
|
| 219 |
+
"rewards/margins": -0.06102731078863144,
|
| 220 |
+
"rewards/rejected": -0.40165242552757263,
|
| 221 |
+
"step": 120
|
| 222 |
+
},
|
| 223 |
+
{
|
| 224 |
+
"epoch": 0.20404159309397685,
|
| 225 |
+
"grad_norm": 0.10464702546596527,
|
| 226 |
+
"learning_rate": 4.907906416994146e-06,
|
| 227 |
+
"logits/chosen": 16.163082122802734,
|
| 228 |
+
"logits/rejected": 16.158031463623047,
|
| 229 |
+
"logps/chosen": -0.3138599991798401,
|
| 230 |
+
"logps/rejected": -0.28097471594810486,
|
| 231 |
+
"loss": 1.0169,
|
| 232 |
+
"rewards/accuracies": 0.30000001192092896,
|
| 233 |
+
"rewards/chosen": -0.47078999876976013,
|
| 234 |
+
"rewards/margins": -0.04932791367173195,
|
| 235 |
+
"rewards/rejected": -0.4214620590209961,
|
| 236 |
+
"step": 130
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 0.219737100255052,
|
| 240 |
+
"grad_norm": 0.16971275210380554,
|
| 241 |
+
"learning_rate": 4.893298743830168e-06,
|
| 242 |
+
"logits/chosen": 16.28864860534668,
|
| 243 |
+
"logits/rejected": 16.151805877685547,
|
| 244 |
+
"logps/chosen": -0.3283368945121765,
|
| 245 |
+
"logps/rejected": -0.2850198745727539,
|
| 246 |
+
"loss": 0.9964,
|
| 247 |
+
"rewards/accuracies": 0.3499999940395355,
|
| 248 |
+
"rewards/chosen": -0.49250537157058716,
|
| 249 |
+
"rewards/margins": -0.06497551500797272,
|
| 250 |
+
"rewards/rejected": -0.42752987146377563,
|
| 251 |
+
"step": 140
|
| 252 |
+
},
|
| 253 |
+
{
|
| 254 |
+
"epoch": 0.23543260741612712,
|
| 255 |
+
"grad_norm": 0.18377964198589325,
|
| 256 |
+
"learning_rate": 4.8776412907378845e-06,
|
| 257 |
+
"logits/chosen": 16.890087127685547,
|
| 258 |
+
"logits/rejected": 16.42388153076172,
|
| 259 |
+
"logps/chosen": -0.33256903290748596,
|
| 260 |
+
"logps/rejected": -0.2939595878124237,
|
| 261 |
+
"loss": 1.0073,
|
| 262 |
+
"rewards/accuracies": 0.30000001192092896,
|
| 263 |
+
"rewards/chosen": -0.49885353446006775,
|
| 264 |
+
"rewards/margins": -0.057914119213819504,
|
| 265 |
+
"rewards/rejected": -0.44093936681747437,
|
| 266 |
+
"step": 150
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 0.23543260741612712,
|
| 270 |
+
"eval_logits/chosen": 16.833438873291016,
|
| 271 |
+
"eval_logits/rejected": 16.328977584838867,
|
| 272 |
+
"eval_logps/chosen": -0.32567569613456726,
|
| 273 |
+
"eval_logps/rejected": -0.35700783133506775,
|
| 274 |
+
"eval_loss": 0.9802881479263306,
|
| 275 |
+
"eval_rewards/accuracies": 0.42307692766189575,
|
| 276 |
+
"eval_rewards/chosen": -0.4885135293006897,
|
| 277 |
+
"eval_rewards/margins": 0.04699822515249252,
|
| 278 |
+
"eval_rewards/rejected": -0.5355117321014404,
|
| 279 |
+
"eval_runtime": 14.5005,
|
| 280 |
+
"eval_samples_per_second": 28.413,
|
| 281 |
+
"eval_steps_per_second": 3.586,
|
| 282 |
+
"step": 150
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 0.2511281145772023,
|
| 286 |
+
"grad_norm": 0.12049826234579086,
|
| 287 |
+
"learning_rate": 4.860940925593703e-06,
|
| 288 |
+
"logits/chosen": 16.505878448486328,
|
| 289 |
+
"logits/rejected": 16.178979873657227,
|
| 290 |
+
"logps/chosen": -0.3397011458873749,
|
| 291 |
+
"logps/rejected": -0.35640352964401245,
|
| 292 |
+
"loss": 0.9795,
|
| 293 |
+
"rewards/accuracies": 0.4124999940395355,
|
| 294 |
+
"rewards/chosen": -0.5095517039299011,
|
| 295 |
+
"rewards/margins": 0.0250535998493433,
|
| 296 |
+
"rewards/rejected": -0.5346053242683411,
|
| 297 |
+
"step": 160
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 0.2668236217382774,
|
| 301 |
+
"grad_norm": 0.09485407918691635,
|
| 302 |
+
"learning_rate": 4.84320497372973e-06,
|
| 303 |
+
"logits/chosen": 16.245588302612305,
|
| 304 |
+
"logits/rejected": 15.922958374023438,
|
| 305 |
+
"logps/chosen": -0.29733315110206604,
|
| 306 |
+
"logps/rejected": -0.3461209237575531,
|
| 307 |
+
"loss": 0.9694,
|
| 308 |
+
"rewards/accuracies": 0.44999998807907104,
|
| 309 |
+
"rewards/chosen": -0.44599977135658264,
|
| 310 |
+
"rewards/margins": 0.0731816366314888,
|
| 311 |
+
"rewards/rejected": -0.5191814303398132,
|
| 312 |
+
"step": 170
|
| 313 |
+
},
|
| 314 |
+
{
|
| 315 |
+
"epoch": 0.28251912889935255,
|
| 316 |
+
"grad_norm": 0.155483216047287,
|
| 317 |
+
"learning_rate": 4.824441214720629e-06,
|
| 318 |
+
"logits/chosen": 16.339645385742188,
|
| 319 |
+
"logits/rejected": 16.115110397338867,
|
| 320 |
+
"logps/chosen": -0.3076801002025604,
|
| 321 |
+
"logps/rejected": -0.3655286729335785,
|
| 322 |
+
"loss": 0.9488,
|
| 323 |
+
"rewards/accuracies": 0.4625000059604645,
|
| 324 |
+
"rewards/chosen": -0.4615201950073242,
|
| 325 |
+
"rewards/margins": 0.0867728441953659,
|
| 326 |
+
"rewards/rejected": -0.5482929944992065,
|
| 327 |
+
"step": 180
|
| 328 |
+
},
|
| 329 |
+
{
|
| 330 |
+
"epoch": 0.2982146360604277,
|
| 331 |
+
"grad_norm": 0.21345795691013336,
|
| 332 |
+
"learning_rate": 4.804657878971252e-06,
|
| 333 |
+
"logits/chosen": 16.491886138916016,
|
| 334 |
+
"logits/rejected": 16.376684188842773,
|
| 335 |
+
"logps/chosen": -0.32437095046043396,
|
| 336 |
+
"logps/rejected": -0.39715105295181274,
|
| 337 |
+
"loss": 0.9423,
|
| 338 |
+
"rewards/accuracies": 0.512499988079071,
|
| 339 |
+
"rewards/chosen": -0.4865564703941345,
|
| 340 |
+
"rewards/margins": 0.10917013883590698,
|
| 341 |
+
"rewards/rejected": -0.5957266092300415,
|
| 342 |
+
"step": 190
|
| 343 |
+
},
|
| 344 |
+
{
|
| 345 |
+
"epoch": 0.31391014322150285,
|
| 346 |
+
"grad_norm": 0.17633090913295746,
|
| 347 |
+
"learning_rate": 4.783863644106502e-06,
|
| 348 |
+
"logits/chosen": 16.6339168548584,
|
| 349 |
+
"logits/rejected": 16.79404640197754,
|
| 350 |
+
"logps/chosen": -0.33018192648887634,
|
| 351 |
+
"logps/rejected": -0.384196400642395,
|
| 352 |
+
"loss": 0.939,
|
| 353 |
+
"rewards/accuracies": 0.42500001192092896,
|
| 354 |
+
"rewards/chosen": -0.4952728748321533,
|
| 355 |
+
"rewards/margins": 0.0810217633843422,
|
| 356 |
+
"rewards/rejected": -0.5762946009635925,
|
| 357 |
+
"step": 200
|
| 358 |
+
},
|
| 359 |
+
{
|
| 360 |
+
"epoch": 0.31391014322150285,
|
| 361 |
+
"eval_logits/chosen": 17.17803192138672,
|
| 362 |
+
"eval_logits/rejected": 16.59328269958496,
|
| 363 |
+
"eval_logps/chosen": -0.33600664138793945,
|
| 364 |
+
"eval_logps/rejected": -0.47861453890800476,
|
| 365 |
+
"eval_loss": 0.9303967356681824,
|
| 366 |
+
"eval_rewards/accuracies": 0.4615384638309479,
|
| 367 |
+
"eval_rewards/chosen": -0.5040098428726196,
|
| 368 |
+
"eval_rewards/margins": 0.21391186118125916,
|
| 369 |
+
"eval_rewards/rejected": -0.717921793460846,
|
| 370 |
+
"eval_runtime": 14.5,
|
| 371 |
+
"eval_samples_per_second": 28.414,
|
| 372 |
+
"eval_steps_per_second": 3.586,
|
| 373 |
+
"step": 200
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.329605650382578,
|
| 377 |
+
"grad_norm": 0.1562221795320511,
|
| 378 |
+
"learning_rate": 4.762067631165049e-06,
|
| 379 |
+
"logits/chosen": 16.597665786743164,
|
| 380 |
+
"logits/rejected": 16.3507022857666,
|
| 381 |
+
"logps/chosen": -0.33709320425987244,
|
| 382 |
+
"logps/rejected": -0.43131333589553833,
|
| 383 |
+
"loss": 0.9107,
|
| 384 |
+
"rewards/accuracies": 0.4749999940395355,
|
| 385 |
+
"rewards/chosen": -0.5056397914886475,
|
| 386 |
+
"rewards/margins": 0.14133022725582123,
|
| 387 |
+
"rewards/rejected": -0.6469700932502747,
|
| 388 |
+
"step": 210
|
| 389 |
+
},
|
| 390 |
+
{
|
| 391 |
+
"epoch": 0.34530115754365315,
|
| 392 |
+
"grad_norm": 0.17680124938488007,
|
| 393 |
+
"learning_rate": 4.7392794005985324e-06,
|
| 394 |
+
"logits/chosen": 16.873676300048828,
|
| 395 |
+
"logits/rejected": 16.61945152282715,
|
| 396 |
+
"logps/chosen": -0.35083168745040894,
|
| 397 |
+
"logps/rejected": -0.537697434425354,
|
| 398 |
+
"loss": 0.9094,
|
| 399 |
+
"rewards/accuracies": 0.5249999761581421,
|
| 400 |
+
"rewards/chosen": -0.526247501373291,
|
| 401 |
+
"rewards/margins": 0.2802986204624176,
|
| 402 |
+
"rewards/rejected": -0.8065462112426758,
|
| 403 |
+
"step": 220
|
| 404 |
+
},
|
| 405 |
+
{
|
| 406 |
+
"epoch": 0.3609966647047283,
|
| 407 |
+
"grad_norm": 0.2662070393562317,
|
| 408 |
+
"learning_rate": 4.715508948078037e-06,
|
| 409 |
+
"logits/chosen": 16.803974151611328,
|
| 410 |
+
"logits/rejected": 16.698320388793945,
|
| 411 |
+
"logps/chosen": -0.3500753343105316,
|
| 412 |
+
"logps/rejected": -0.5399882793426514,
|
| 413 |
+
"loss": 0.8871,
|
| 414 |
+
"rewards/accuracies": 0.5249999761581421,
|
| 415 |
+
"rewards/chosen": -0.5251129865646362,
|
| 416 |
+
"rewards/margins": 0.2848694324493408,
|
| 417 |
+
"rewards/rejected": -0.8099824786186218,
|
| 418 |
+
"step": 230
|
| 419 |
+
},
|
| 420 |
+
{
|
| 421 |
+
"epoch": 0.3766921718658034,
|
| 422 |
+
"grad_norm": 0.17161667346954346,
|
| 423 |
+
"learning_rate": 4.690766700109659e-06,
|
| 424 |
+
"logits/chosen": 16.67904281616211,
|
| 425 |
+
"logits/rejected": 16.410579681396484,
|
| 426 |
+
"logps/chosen": -0.33549198508262634,
|
| 427 |
+
"logps/rejected": -0.4875665605068207,
|
| 428 |
+
"loss": 0.9037,
|
| 429 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 430 |
+
"rewards/chosen": -0.5032380223274231,
|
| 431 |
+
"rewards/margins": 0.22811183333396912,
|
| 432 |
+
"rewards/rejected": -0.7313498258590698,
|
| 433 |
+
"step": 240
|
| 434 |
+
},
|
| 435 |
+
{
|
| 436 |
+
"epoch": 0.39238767902687854,
|
| 437 |
+
"grad_norm": 1.106021761894226,
|
| 438 |
+
"learning_rate": 4.665063509461098e-06,
|
| 439 |
+
"logits/chosen": 17.208127975463867,
|
| 440 |
+
"logits/rejected": 16.654085159301758,
|
| 441 |
+
"logps/chosen": -0.3752726912498474,
|
| 442 |
+
"logps/rejected": -0.4947708249092102,
|
| 443 |
+
"loss": 0.8606,
|
| 444 |
+
"rewards/accuracies": 0.4375,
|
| 445 |
+
"rewards/chosen": -0.5629090070724487,
|
| 446 |
+
"rewards/margins": 0.1792471706867218,
|
| 447 |
+
"rewards/rejected": -0.7421562671661377,
|
| 448 |
+
"step": 250
|
| 449 |
+
},
|
| 450 |
+
{
|
| 451 |
+
"epoch": 0.39238767902687854,
|
| 452 |
+
"eval_logits/chosen": 17.32963752746582,
|
| 453 |
+
"eval_logits/rejected": 16.589412689208984,
|
| 454 |
+
"eval_logps/chosen": -0.37825876474380493,
|
| 455 |
+
"eval_logps/rejected": -0.9001243114471436,
|
| 456 |
+
"eval_loss": 0.8168494701385498,
|
| 457 |
+
"eval_rewards/accuracies": 0.5192307829856873,
|
| 458 |
+
"eval_rewards/chosen": -0.5673881769180298,
|
| 459 |
+
"eval_rewards/margins": 0.7827982306480408,
|
| 460 |
+
"eval_rewards/rejected": -1.3501865863800049,
|
| 461 |
+
"eval_runtime": 14.5053,
|
| 462 |
+
"eval_samples_per_second": 28.403,
|
| 463 |
+
"eval_steps_per_second": 3.585,
|
| 464 |
+
"step": 250
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.4080831861879537,
|
| 468 |
+
"grad_norm": 0.29772138595581055,
|
| 469 |
+
"learning_rate": 4.638410650401267e-06,
|
| 470 |
+
"logits/chosen": 16.713775634765625,
|
| 471 |
+
"logits/rejected": 16.35211944580078,
|
| 472 |
+
"logps/chosen": -0.3877524733543396,
|
| 473 |
+
"logps/rejected": -0.8511163592338562,
|
| 474 |
+
"loss": 0.8138,
|
| 475 |
+
"rewards/accuracies": 0.44999998807907104,
|
| 476 |
+
"rewards/chosen": -0.581628680229187,
|
| 477 |
+
"rewards/margins": 0.6950457692146301,
|
| 478 |
+
"rewards/rejected": -1.276674509048462,
|
| 479 |
+
"step": 260
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 0.42377869334902885,
|
| 483 |
+
"grad_norm": 0.31860050559043884,
|
| 484 |
+
"learning_rate": 4.610819813755038e-06,
|
| 485 |
+
"logits/chosen": 17.09469985961914,
|
| 486 |
+
"logits/rejected": 16.5472412109375,
|
| 487 |
+
"logps/chosen": -0.47504258155822754,
|
| 488 |
+
"logps/rejected": -1.3266533613204956,
|
| 489 |
+
"loss": 0.7318,
|
| 490 |
+
"rewards/accuracies": 0.512499988079071,
|
| 491 |
+
"rewards/chosen": -0.7125638723373413,
|
| 492 |
+
"rewards/margins": 1.2774161100387573,
|
| 493 |
+
"rewards/rejected": -1.9899799823760986,
|
| 494 |
+
"step": 270
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 0.439474200510104,
|
| 498 |
+
"grad_norm": 0.6508163809776306,
|
| 499 |
+
"learning_rate": 4.582303101775249e-06,
|
| 500 |
+
"logits/chosen": 17.11003303527832,
|
| 501 |
+
"logits/rejected": 16.6564998626709,
|
| 502 |
+
"logps/chosen": -0.49357643723487854,
|
| 503 |
+
"logps/rejected": -1.4481580257415771,
|
| 504 |
+
"loss": 0.7599,
|
| 505 |
+
"rewards/accuracies": 0.6625000238418579,
|
| 506 |
+
"rewards/chosen": -0.7403645515441895,
|
| 507 |
+
"rewards/margins": 1.4318726062774658,
|
| 508 |
+
"rewards/rejected": -2.1722371578216553,
|
| 509 |
+
"step": 280
|
| 510 |
+
},
|
| 511 |
+
{
|
| 512 |
+
"epoch": 0.45516970767117915,
|
| 513 |
+
"grad_norm": 0.32430580258369446,
|
| 514 |
+
"learning_rate": 4.55287302283426e-06,
|
| 515 |
+
"logits/chosen": 16.830989837646484,
|
| 516 |
+
"logits/rejected": 16.384944915771484,
|
| 517 |
+
"logps/chosen": -0.593712329864502,
|
| 518 |
+
"logps/rejected": -1.7630856037139893,
|
| 519 |
+
"loss": 0.7336,
|
| 520 |
+
"rewards/accuracies": 0.675000011920929,
|
| 521 |
+
"rewards/chosen": -0.8905684351921082,
|
| 522 |
+
"rewards/margins": 1.7540600299835205,
|
| 523 |
+
"rewards/rejected": -2.6446282863616943,
|
| 524 |
+
"step": 290
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 0.47086521483225424,
|
| 528 |
+
"grad_norm": 0.8555932641029358,
|
| 529 |
+
"learning_rate": 4.522542485937369e-06,
|
| 530 |
+
"logits/chosen": 16.72231674194336,
|
| 531 |
+
"logits/rejected": 16.28726577758789,
|
| 532 |
+
"logps/chosen": -0.5670709609985352,
|
| 533 |
+
"logps/rejected": -2.0420775413513184,
|
| 534 |
+
"loss": 0.6861,
|
| 535 |
+
"rewards/accuracies": 0.75,
|
| 536 |
+
"rewards/chosen": -0.8506065607070923,
|
| 537 |
+
"rewards/margins": 2.2125096321105957,
|
| 538 |
+
"rewards/rejected": -3.0631160736083984,
|
| 539 |
+
"step": 300
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"epoch": 0.47086521483225424,
|
| 543 |
+
"eval_logits/chosen": 17.13121795654297,
|
| 544 |
+
"eval_logits/rejected": 16.268341064453125,
|
| 545 |
+
"eval_logps/chosen": -0.6842947602272034,
|
| 546 |
+
"eval_logps/rejected": -2.119321584701538,
|
| 547 |
+
"eval_loss": 0.7583853602409363,
|
| 548 |
+
"eval_rewards/accuracies": 0.75,
|
| 549 |
+
"eval_rewards/chosen": -1.026442289352417,
|
| 550 |
+
"eval_rewards/margins": 2.1525399684906006,
|
| 551 |
+
"eval_rewards/rejected": -3.1789822578430176,
|
| 552 |
+
"eval_runtime": 14.5007,
|
| 553 |
+
"eval_samples_per_second": 28.412,
|
| 554 |
+
"eval_steps_per_second": 3.586,
|
| 555 |
+
"step": 300
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.4865607219933294,
|
| 559 |
+
"grad_norm": 0.6059070825576782,
|
| 560 |
+
"learning_rate": 4.491324795060491e-06,
|
| 561 |
+
"logits/chosen": 17.147808074951172,
|
| 562 |
+
"logits/rejected": 16.194652557373047,
|
| 563 |
+
"logps/chosen": -0.8036400079727173,
|
| 564 |
+
"logps/rejected": -2.289825201034546,
|
| 565 |
+
"loss": 0.7163,
|
| 566 |
+
"rewards/accuracies": 0.6875,
|
| 567 |
+
"rewards/chosen": -1.2054599523544312,
|
| 568 |
+
"rewards/margins": 2.229278087615967,
|
| 569 |
+
"rewards/rejected": -3.4347376823425293,
|
| 570 |
+
"step": 310
|
| 571 |
+
},
|
| 572 |
+
{
|
| 573 |
+
"epoch": 0.5022562291544046,
|
| 574 |
+
"grad_norm": 1.8073927164077759,
|
| 575 |
+
"learning_rate": 4.4592336433146e-06,
|
| 576 |
+
"logits/chosen": 16.94902992248535,
|
| 577 |
+
"logits/rejected": 16.066068649291992,
|
| 578 |
+
"logps/chosen": -1.198162317276001,
|
| 579 |
+
"logps/rejected": -2.2922632694244385,
|
| 580 |
+
"loss": 0.6894,
|
| 581 |
+
"rewards/accuracies": 0.675000011920929,
|
| 582 |
+
"rewards/chosen": -1.7972434759140015,
|
| 583 |
+
"rewards/margins": 1.6411516666412354,
|
| 584 |
+
"rewards/rejected": -3.4383950233459473,
|
| 585 |
+
"step": 320
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 0.5179517363154797,
|
| 589 |
+
"grad_norm": 3.746042490005493,
|
| 590 |
+
"learning_rate": 4.426283106939474e-06,
|
| 591 |
+
"logits/chosen": 16.31036376953125,
|
| 592 |
+
"logits/rejected": 15.991762161254883,
|
| 593 |
+
"logps/chosen": -1.6245781183242798,
|
| 594 |
+
"logps/rejected": -2.553597927093506,
|
| 595 |
+
"loss": 0.6607,
|
| 596 |
+
"rewards/accuracies": 0.7875000238418579,
|
| 597 |
+
"rewards/chosen": -2.4368672370910645,
|
| 598 |
+
"rewards/margins": 1.3935294151306152,
|
| 599 |
+
"rewards/rejected": -3.830397129058838,
|
| 600 |
+
"step": 330
|
| 601 |
+
},
|
| 602 |
+
{
|
| 603 |
+
"epoch": 0.5336472434765548,
|
| 604 |
+
"grad_norm": 2.098111867904663,
|
| 605 |
+
"learning_rate": 4.3924876391293915e-06,
|
| 606 |
+
"logits/chosen": 16.59554672241211,
|
| 607 |
+
"logits/rejected": 15.915553092956543,
|
| 608 |
+
"logps/chosen": -2.183227062225342,
|
| 609 |
+
"logps/rejected": -3.4911434650421143,
|
| 610 |
+
"loss": 0.6381,
|
| 611 |
+
"rewards/accuracies": 0.8500000238418579,
|
| 612 |
+
"rewards/chosen": -3.274840831756592,
|
| 613 |
+
"rewards/margins": 1.9618743658065796,
|
| 614 |
+
"rewards/rejected": -5.236715316772461,
|
| 615 |
+
"step": 340
|
| 616 |
+
},
|
| 617 |
+
{
|
| 618 |
+
"epoch": 0.5493427506376299,
|
| 619 |
+
"grad_norm": 2.153958320617676,
|
| 620 |
+
"learning_rate": 4.357862063693486e-06,
|
| 621 |
+
"logits/chosen": 16.332544326782227,
|
| 622 |
+
"logits/rejected": 15.691922187805176,
|
| 623 |
+
"logps/chosen": -2.673710346221924,
|
| 624 |
+
"logps/rejected": -3.8687057495117188,
|
| 625 |
+
"loss": 0.5752,
|
| 626 |
+
"rewards/accuracies": 0.862500011920929,
|
| 627 |
+
"rewards/chosen": -4.010565757751465,
|
| 628 |
+
"rewards/margins": 1.792493224143982,
|
| 629 |
+
"rewards/rejected": -5.8030595779418945,
|
| 630 |
+
"step": 350
|
| 631 |
+
},
|
| 632 |
+
{
|
| 633 |
+
"epoch": 0.5493427506376299,
|
| 634 |
+
"eval_logits/chosen": 16.272428512573242,
|
| 635 |
+
"eval_logits/rejected": 15.381678581237793,
|
| 636 |
+
"eval_logps/chosen": -3.0390572547912598,
|
| 637 |
+
"eval_logps/rejected": -4.695068836212158,
|
| 638 |
+
"eval_loss": 0.5928590893745422,
|
| 639 |
+
"eval_rewards/accuracies": 0.9230769276618958,
|
| 640 |
+
"eval_rewards/chosen": -4.5585856437683105,
|
| 641 |
+
"eval_rewards/margins": 2.4840168952941895,
|
| 642 |
+
"eval_rewards/rejected": -7.0426025390625,
|
| 643 |
+
"eval_runtime": 14.5008,
|
| 644 |
+
"eval_samples_per_second": 28.412,
|
| 645 |
+
"eval_steps_per_second": 3.586,
|
| 646 |
+
"step": 350
|
| 647 |
+
},
|
| 648 |
+
{
|
| 649 |
+
"epoch": 0.5650382577987051,
|
| 650 |
+
"grad_norm": 1.5755672454833984,
|
| 651 |
+
"learning_rate": 4.322421568553529e-06,
|
| 652 |
+
"logits/chosen": 15.589811325073242,
|
| 653 |
+
"logits/rejected": 14.780921936035156,
|
| 654 |
+
"logps/chosen": -3.069565534591675,
|
| 655 |
+
"logps/rejected": -4.581957817077637,
|
| 656 |
+
"loss": 0.5275,
|
| 657 |
+
"rewards/accuracies": 0.8125,
|
| 658 |
+
"rewards/chosen": -4.604348182678223,
|
| 659 |
+
"rewards/margins": 2.268587827682495,
|
| 660 |
+
"rewards/rejected": -6.872936248779297,
|
| 661 |
+
"step": 360
|
| 662 |
+
},
|
| 663 |
+
{
|
| 664 |
+
"epoch": 0.5807337649597802,
|
| 665 |
+
"grad_norm": 1.8776415586471558,
|
| 666 |
+
"learning_rate": 4.286181699082008e-06,
|
| 667 |
+
"logits/chosen": 15.172673225402832,
|
| 668 |
+
"logits/rejected": 14.400335311889648,
|
| 669 |
+
"logps/chosen": -3.9551639556884766,
|
| 670 |
+
"logps/rejected": -5.95252799987793,
|
| 671 |
+
"loss": 0.5113,
|
| 672 |
+
"rewards/accuracies": 0.887499988079071,
|
| 673 |
+
"rewards/chosen": -5.932745933532715,
|
| 674 |
+
"rewards/margins": 2.996046543121338,
|
| 675 |
+
"rewards/rejected": -8.928792953491211,
|
| 676 |
+
"step": 370
|
| 677 |
+
},
|
| 678 |
+
{
|
| 679 |
+
"epoch": 0.5964292721208554,
|
| 680 |
+
"grad_norm": 1.5507289171218872,
|
| 681 |
+
"learning_rate": 4.249158351283414e-06,
|
| 682 |
+
"logits/chosen": 14.8423433303833,
|
| 683 |
+
"logits/rejected": 14.592633247375488,
|
| 684 |
+
"logps/chosen": -3.7946255207061768,
|
| 685 |
+
"logps/rejected": -5.3750481605529785,
|
| 686 |
+
"loss": 0.4862,
|
| 687 |
+
"rewards/accuracies": 0.887499988079071,
|
| 688 |
+
"rewards/chosen": -5.691938877105713,
|
| 689 |
+
"rewards/margins": 2.370633125305176,
|
| 690 |
+
"rewards/rejected": -8.062570571899414,
|
| 691 |
+
"step": 380
|
| 692 |
+
},
|
| 693 |
+
{
|
| 694 |
+
"epoch": 0.6121247792819305,
|
| 695 |
+
"grad_norm": 3.4510324001312256,
|
| 696 |
+
"learning_rate": 4.211367764821722e-06,
|
| 697 |
+
"logits/chosen": 14.586801528930664,
|
| 698 |
+
"logits/rejected": 14.031987190246582,
|
| 699 |
+
"logps/chosen": -4.391470432281494,
|
| 700 |
+
"logps/rejected": -6.2107672691345215,
|
| 701 |
+
"loss": 0.5117,
|
| 702 |
+
"rewards/accuracies": 0.8500000238418579,
|
| 703 |
+
"rewards/chosen": -6.587205410003662,
|
| 704 |
+
"rewards/margins": 2.72894549369812,
|
| 705 |
+
"rewards/rejected": -9.316150665283203,
|
| 706 |
+
"step": 390
|
| 707 |
+
},
|
| 708 |
+
{
|
| 709 |
+
"epoch": 0.6278202864430057,
|
| 710 |
+
"grad_norm": 3.3816750049591064,
|
| 711 |
+
"learning_rate": 4.172826515897146e-06,
|
| 712 |
+
"logits/chosen": 14.523185729980469,
|
| 713 |
+
"logits/rejected": 14.266815185546875,
|
| 714 |
+
"logps/chosen": -4.090173244476318,
|
| 715 |
+
"logps/rejected": -5.582955837249756,
|
| 716 |
+
"loss": 0.5182,
|
| 717 |
+
"rewards/accuracies": 0.762499988079071,
|
| 718 |
+
"rewards/chosen": -6.135260581970215,
|
| 719 |
+
"rewards/margins": 2.2391738891601562,
|
| 720 |
+
"rewards/rejected": -8.374434471130371,
|
| 721 |
+
"step": 400
|
| 722 |
+
},
|
| 723 |
+
{
|
| 724 |
+
"epoch": 0.6278202864430057,
|
| 725 |
+
"eval_logits/chosen": 15.142684936523438,
|
| 726 |
+
"eval_logits/rejected": 14.101082801818848,
|
| 727 |
+
"eval_logps/chosen": -3.866588830947876,
|
| 728 |
+
"eval_logps/rejected": -6.100707530975342,
|
| 729 |
+
"eval_loss": 0.48699691891670227,
|
| 730 |
+
"eval_rewards/accuracies": 0.9615384340286255,
|
| 731 |
+
"eval_rewards/chosen": -5.7998833656311035,
|
| 732 |
+
"eval_rewards/margins": 3.351177215576172,
|
| 733 |
+
"eval_rewards/rejected": -9.15106201171875,
|
| 734 |
+
"eval_runtime": 14.5024,
|
| 735 |
+
"eval_samples_per_second": 28.409,
|
| 736 |
+
"eval_steps_per_second": 3.586,
|
| 737 |
+
"step": 400
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 0.6435157936040808,
|
| 741 |
+
"grad_norm": 1.845595359802246,
|
| 742 |
+
"learning_rate": 4.133551509975264e-06,
|
| 743 |
+
"logits/chosen": 14.54790210723877,
|
| 744 |
+
"logits/rejected": 13.707855224609375,
|
| 745 |
+
"logps/chosen": -4.112462043762207,
|
| 746 |
+
"logps/rejected": -6.583975315093994,
|
| 747 |
+
"loss": 0.5107,
|
| 748 |
+
"rewards/accuracies": 0.800000011920929,
|
| 749 |
+
"rewards/chosen": -6.168692111968994,
|
| 750 |
+
"rewards/margins": 3.707270383834839,
|
| 751 |
+
"rewards/rejected": -9.875962257385254,
|
| 752 |
+
"step": 410
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 0.659211300765156,
|
| 756 |
+
"grad_norm": 2.152916193008423,
|
| 757 |
+
"learning_rate": 4.093559974371725e-06,
|
| 758 |
+
"logits/chosen": 14.153576850891113,
|
| 759 |
+
"logits/rejected": 13.553201675415039,
|
| 760 |
+
"logps/chosen": -4.3408918380737305,
|
| 761 |
+
"logps/rejected": -6.828585147857666,
|
| 762 |
+
"loss": 0.41,
|
| 763 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 764 |
+
"rewards/chosen": -6.511338233947754,
|
| 765 |
+
"rewards/margins": 3.7315402030944824,
|
| 766 |
+
"rewards/rejected": -10.242877960205078,
|
| 767 |
+
"step": 420
|
| 768 |
+
},
|
| 769 |
+
{
|
| 770 |
+
"epoch": 0.6749068079262311,
|
| 771 |
+
"grad_norm": 2.305433750152588,
|
| 772 |
+
"learning_rate": 4.052869450695776e-06,
|
| 773 |
+
"logits/chosen": 13.992170333862305,
|
| 774 |
+
"logits/rejected": 13.264738082885742,
|
| 775 |
+
"logps/chosen": -4.259932518005371,
|
| 776 |
+
"logps/rejected": -6.089646339416504,
|
| 777 |
+
"loss": 0.4122,
|
| 778 |
+
"rewards/accuracies": 0.862500011920929,
|
| 779 |
+
"rewards/chosen": -6.389898777008057,
|
| 780 |
+
"rewards/margins": 2.7445709705352783,
|
| 781 |
+
"rewards/rejected": -9.134469985961914,
|
| 782 |
+
"step": 430
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 0.6906023150873063,
|
| 786 |
+
"grad_norm": 2.397674322128296,
|
| 787 |
+
"learning_rate": 4.011497787155938e-06,
|
| 788 |
+
"logits/chosen": 13.975759506225586,
|
| 789 |
+
"logits/rejected": 13.173799514770508,
|
| 790 |
+
"logps/chosen": -4.2483391761779785,
|
| 791 |
+
"logps/rejected": -6.487355709075928,
|
| 792 |
+
"loss": 0.4397,
|
| 793 |
+
"rewards/accuracies": 0.8125,
|
| 794 |
+
"rewards/chosen": -6.3725080490112305,
|
| 795 |
+
"rewards/margins": 3.3585267066955566,
|
| 796 |
+
"rewards/rejected": -9.731034278869629,
|
| 797 |
+
"step": 440
|
| 798 |
+
},
|
| 799 |
+
{
|
| 800 |
+
"epoch": 0.7062978222483814,
|
| 801 |
+
"grad_norm": 2.3987627029418945,
|
| 802 |
+
"learning_rate": 3.969463130731183e-06,
|
| 803 |
+
"logits/chosen": 13.777560234069824,
|
| 804 |
+
"logits/rejected": 12.936296463012695,
|
| 805 |
+
"logps/chosen": -3.86059832572937,
|
| 806 |
+
"logps/rejected": -6.495786190032959,
|
| 807 |
+
"loss": 0.444,
|
| 808 |
+
"rewards/accuracies": 0.862500011920929,
|
| 809 |
+
"rewards/chosen": -5.790897846221924,
|
| 810 |
+
"rewards/margins": 3.952782392501831,
|
| 811 |
+
"rewards/rejected": -9.743680000305176,
|
| 812 |
+
"step": 450
|
| 813 |
+
},
|
| 814 |
+
{
|
| 815 |
+
"epoch": 0.7062978222483814,
|
| 816 |
+
"eval_logits/chosen": 14.397418975830078,
|
| 817 |
+
"eval_logits/rejected": 13.271268844604492,
|
| 818 |
+
"eval_logps/chosen": -3.9544241428375244,
|
| 819 |
+
"eval_logps/rejected": -6.69989013671875,
|
| 820 |
+
"eval_loss": 0.435256689786911,
|
| 821 |
+
"eval_rewards/accuracies": 0.9615384340286255,
|
| 822 |
+
"eval_rewards/chosen": -5.931635856628418,
|
| 823 |
+
"eval_rewards/margins": 4.118198871612549,
|
| 824 |
+
"eval_rewards/rejected": -10.049835205078125,
|
| 825 |
+
"eval_runtime": 14.5029,
|
| 826 |
+
"eval_samples_per_second": 28.408,
|
| 827 |
+
"eval_steps_per_second": 3.585,
|
| 828 |
+
"step": 450
|
| 829 |
+
},
|
| 830 |
+
{
|
| 831 |
+
"epoch": 0.7219933294094566,
|
| 832 |
+
"grad_norm": 2.3673465251922607,
|
| 833 |
+
"learning_rate": 3.92678391921108e-06,
|
| 834 |
+
"logits/chosen": 13.67004680633545,
|
| 835 |
+
"logits/rejected": 12.689542770385742,
|
| 836 |
+
"logps/chosen": -4.037893295288086,
|
| 837 |
+
"logps/rejected": -6.727287292480469,
|
| 838 |
+
"loss": 0.3858,
|
| 839 |
+
"rewards/accuracies": 0.925000011920929,
|
| 840 |
+
"rewards/chosen": -6.056839942932129,
|
| 841 |
+
"rewards/margins": 4.034090995788574,
|
| 842 |
+
"rewards/rejected": -10.09093189239502,
|
| 843 |
+
"step": 460
|
| 844 |
+
},
|
| 845 |
+
{
|
| 846 |
+
"epoch": 0.7376888365705316,
|
| 847 |
+
"grad_norm": 2.9868297576904297,
|
| 848 |
+
"learning_rate": 3.88347887310836e-06,
|
| 849 |
+
"logits/chosen": 13.124593734741211,
|
| 850 |
+
"logits/rejected": 12.576837539672852,
|
| 851 |
+
"logps/chosen": -4.324252605438232,
|
| 852 |
+
"logps/rejected": -6.761715888977051,
|
| 853 |
+
"loss": 0.363,
|
| 854 |
+
"rewards/accuracies": 0.862500011920929,
|
| 855 |
+
"rewards/chosen": -6.486379146575928,
|
| 856 |
+
"rewards/margins": 3.6561942100524902,
|
| 857 |
+
"rewards/rejected": -10.142572402954102,
|
| 858 |
+
"step": 470
|
| 859 |
+
},
|
| 860 |
+
{
|
| 861 |
+
"epoch": 0.7533843437316068,
|
| 862 |
+
"grad_norm": 2.613318920135498,
|
| 863 |
+
"learning_rate": 3.839566987447492e-06,
|
| 864 |
+
"logits/chosen": 13.514410018920898,
|
| 865 |
+
"logits/rejected": 12.838116645812988,
|
| 866 |
+
"logps/chosen": -4.333093166351318,
|
| 867 |
+
"logps/rejected": -6.704646110534668,
|
| 868 |
+
"loss": 0.3274,
|
| 869 |
+
"rewards/accuracies": 0.925000011920929,
|
| 870 |
+
"rewards/chosen": -6.499639987945557,
|
| 871 |
+
"rewards/margins": 3.5573298931121826,
|
| 872 |
+
"rewards/rejected": -10.056970596313477,
|
| 873 |
+
"step": 480
|
| 874 |
+
},
|
| 875 |
+
{
|
| 876 |
+
"epoch": 0.7690798508926819,
|
| 877 |
+
"grad_norm": 2.1568610668182373,
|
| 878 |
+
"learning_rate": 3.795067523432826e-06,
|
| 879 |
+
"logits/chosen": 13.456674575805664,
|
| 880 |
+
"logits/rejected": 12.37132453918457,
|
| 881 |
+
"logps/chosen": -4.630190849304199,
|
| 882 |
+
"logps/rejected": -7.871635437011719,
|
| 883 |
+
"loss": 0.3822,
|
| 884 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 885 |
+
"rewards/chosen": -6.945285797119141,
|
| 886 |
+
"rewards/margins": 4.862167835235596,
|
| 887 |
+
"rewards/rejected": -11.807454109191895,
|
| 888 |
+
"step": 490
|
| 889 |
+
},
|
| 890 |
+
{
|
| 891 |
+
"epoch": 0.7847753580537571,
|
| 892 |
+
"grad_norm": 2.516496419906616,
|
| 893 |
+
"learning_rate": 3.7500000000000005e-06,
|
| 894 |
+
"logits/chosen": 13.195734024047852,
|
| 895 |
+
"logits/rejected": 12.557401657104492,
|
| 896 |
+
"logps/chosen": -3.9327914714813232,
|
| 897 |
+
"logps/rejected": -6.561183929443359,
|
| 898 |
+
"loss": 0.352,
|
| 899 |
+
"rewards/accuracies": 0.887499988079071,
|
| 900 |
+
"rewards/chosen": -5.899188041687012,
|
| 901 |
+
"rewards/margins": 3.9425880908966064,
|
| 902 |
+
"rewards/rejected": -9.841775894165039,
|
| 903 |
+
"step": 500
|
| 904 |
+
},
|
| 905 |
+
{
|
| 906 |
+
"epoch": 0.7847753580537571,
|
| 907 |
+
"eval_logits/chosen": 13.891414642333984,
|
| 908 |
+
"eval_logits/rejected": 12.734375,
|
| 909 |
+
"eval_logps/chosen": -3.924809455871582,
|
| 910 |
+
"eval_logps/rejected": -7.003909587860107,
|
| 911 |
+
"eval_loss": 0.3888731002807617,
|
| 912 |
+
"eval_rewards/accuracies": 1.0,
|
| 913 |
+
"eval_rewards/chosen": -5.887214183807373,
|
| 914 |
+
"eval_rewards/margins": 4.618649482727051,
|
| 915 |
+
"eval_rewards/rejected": -10.505864143371582,
|
| 916 |
+
"eval_runtime": 14.5029,
|
| 917 |
+
"eval_samples_per_second": 28.408,
|
| 918 |
+
"eval_steps_per_second": 3.585,
|
| 919 |
+
"step": 500
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"epoch": 0.8004708652148322,
|
| 923 |
+
"grad_norm": 5.4663896560668945,
|
| 924 |
+
"learning_rate": 3.7043841852542884e-06,
|
| 925 |
+
"logits/chosen": 12.861404418945312,
|
| 926 |
+
"logits/rejected": 11.9797945022583,
|
| 927 |
+
"logps/chosen": -4.245055198669434,
|
| 928 |
+
"logps/rejected": -7.063906669616699,
|
| 929 |
+
"loss": 0.4009,
|
| 930 |
+
"rewards/accuracies": 0.8500000238418579,
|
| 931 |
+
"rewards/chosen": -6.36758279800415,
|
| 932 |
+
"rewards/margins": 4.228276252746582,
|
| 933 |
+
"rewards/rejected": -10.59585952758789,
|
| 934 |
+
"step": 510
|
| 935 |
+
},
|
| 936 |
+
{
|
| 937 |
+
"epoch": 0.8161663723759074,
|
| 938 |
+
"grad_norm": 2.4403271675109863,
|
| 939 |
+
"learning_rate": 3.658240087799655e-06,
|
| 940 |
+
"logits/chosen": 12.84996223449707,
|
| 941 |
+
"logits/rejected": 12.256246566772461,
|
| 942 |
+
"logps/chosen": -4.2785773277282715,
|
| 943 |
+
"logps/rejected": -7.003395080566406,
|
| 944 |
+
"loss": 0.3327,
|
| 945 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 946 |
+
"rewards/chosen": -6.417865753173828,
|
| 947 |
+
"rewards/margins": 4.0872273445129395,
|
| 948 |
+
"rewards/rejected": -10.505093574523926,
|
| 949 |
+
"step": 520
|
| 950 |
+
},
|
| 951 |
+
{
|
| 952 |
+
"epoch": 0.8318618795369825,
|
| 953 |
+
"grad_norm": 3.596749782562256,
|
| 954 |
+
"learning_rate": 3.611587947962319e-06,
|
| 955 |
+
"logits/chosen": 13.3360595703125,
|
| 956 |
+
"logits/rejected": 12.449459075927734,
|
| 957 |
+
"logps/chosen": -4.8148322105407715,
|
| 958 |
+
"logps/rejected": -7.221930503845215,
|
| 959 |
+
"loss": 0.3255,
|
| 960 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 961 |
+
"rewards/chosen": -7.222248077392578,
|
| 962 |
+
"rewards/margins": 3.610647201538086,
|
| 963 |
+
"rewards/rejected": -10.832895278930664,
|
| 964 |
+
"step": 530
|
| 965 |
+
},
|
| 966 |
+
{
|
| 967 |
+
"epoch": 0.8475573866980577,
|
| 968 |
+
"grad_norm": 4.537969589233398,
|
| 969 |
+
"learning_rate": 3.564448228912682e-06,
|
| 970 |
+
"logits/chosen": 12.63983154296875,
|
| 971 |
+
"logits/rejected": 12.042104721069336,
|
| 972 |
+
"logps/chosen": -4.933573246002197,
|
| 973 |
+
"logps/rejected": -7.656388282775879,
|
| 974 |
+
"loss": 0.2941,
|
| 975 |
+
"rewards/accuracies": 0.925000011920929,
|
| 976 |
+
"rewards/chosen": -7.400360107421875,
|
| 977 |
+
"rewards/margins": 4.084224224090576,
|
| 978 |
+
"rewards/rejected": -11.484583854675293,
|
| 979 |
+
"step": 540
|
| 980 |
+
},
|
| 981 |
+
{
|
| 982 |
+
"epoch": 0.8632528938591328,
|
| 983 |
+
"grad_norm": 2.4124038219451904,
|
| 984 |
+
"learning_rate": 3.516841607689501e-06,
|
| 985 |
+
"logits/chosen": 13.053291320800781,
|
| 986 |
+
"logits/rejected": 12.351530075073242,
|
| 987 |
+
"logps/chosen": -5.196786403656006,
|
| 988 |
+
"logps/rejected": -7.8461012840271,
|
| 989 |
+
"loss": 0.3906,
|
| 990 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 991 |
+
"rewards/chosen": -7.795179843902588,
|
| 992 |
+
"rewards/margins": 3.9739716053009033,
|
| 993 |
+
"rewards/rejected": -11.76915168762207,
|
| 994 |
+
"step": 550
|
| 995 |
+
},
|
| 996 |
+
{
|
| 997 |
+
"epoch": 0.8632528938591328,
|
| 998 |
+
"eval_logits/chosen": 13.596405982971191,
|
| 999 |
+
"eval_logits/rejected": 12.482269287109375,
|
| 1000 |
+
"eval_logps/chosen": -4.328949928283691,
|
| 1001 |
+
"eval_logps/rejected": -7.602427959442139,
|
| 1002 |
+
"eval_loss": 0.3577499985694885,
|
| 1003 |
+
"eval_rewards/accuracies": 0.9807692170143127,
|
| 1004 |
+
"eval_rewards/chosen": -6.493425369262695,
|
| 1005 |
+
"eval_rewards/margins": 4.910217761993408,
|
| 1006 |
+
"eval_rewards/rejected": -11.403642654418945,
|
| 1007 |
+
"eval_runtime": 14.5032,
|
| 1008 |
+
"eval_samples_per_second": 28.408,
|
| 1009 |
+
"eval_steps_per_second": 3.585,
|
| 1010 |
+
"step": 550
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 0.878948401020208,
|
| 1014 |
+
"grad_norm": 1.42177414894104,
|
| 1015 |
+
"learning_rate": 3.4687889661302577e-06,
|
| 1016 |
+
"logits/chosen": 13.275362014770508,
|
| 1017 |
+
"logits/rejected": 12.30308723449707,
|
| 1018 |
+
"logps/chosen": -4.182652950286865,
|
| 1019 |
+
"logps/rejected": -7.163094997406006,
|
| 1020 |
+
"loss": 0.3692,
|
| 1021 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1022 |
+
"rewards/chosen": -6.273979663848877,
|
| 1023 |
+
"rewards/margins": 4.470663547515869,
|
| 1024 |
+
"rewards/rejected": -10.744643211364746,
|
| 1025 |
+
"step": 560
|
| 1026 |
+
},
|
| 1027 |
+
{
|
| 1028 |
+
"epoch": 0.8946439081812831,
|
| 1029 |
+
"grad_norm": 2.431758403778076,
|
| 1030 |
+
"learning_rate": 3.4203113817116955e-06,
|
| 1031 |
+
"logits/chosen": 13.407594680786133,
|
| 1032 |
+
"logits/rejected": 12.078115463256836,
|
| 1033 |
+
"logps/chosen": -4.193841457366943,
|
| 1034 |
+
"logps/rejected": -7.832788944244385,
|
| 1035 |
+
"loss": 0.3487,
|
| 1036 |
+
"rewards/accuracies": 0.9375,
|
| 1037 |
+
"rewards/chosen": -6.290762901306152,
|
| 1038 |
+
"rewards/margins": 5.458420753479004,
|
| 1039 |
+
"rewards/rejected": -11.749183654785156,
|
| 1040 |
+
"step": 570
|
| 1041 |
+
},
|
| 1042 |
+
{
|
| 1043 |
+
"epoch": 0.9103394153423583,
|
| 1044 |
+
"grad_norm": 2.549168348312378,
|
| 1045 |
+
"learning_rate": 3.3714301183045382e-06,
|
| 1046 |
+
"logits/chosen": 12.668423652648926,
|
| 1047 |
+
"logits/rejected": 11.851642608642578,
|
| 1048 |
+
"logps/chosen": -4.831603050231934,
|
| 1049 |
+
"logps/rejected": -7.4626054763793945,
|
| 1050 |
+
"loss": 0.3717,
|
| 1051 |
+
"rewards/accuracies": 0.8500000238418579,
|
| 1052 |
+
"rewards/chosen": -7.2474045753479,
|
| 1053 |
+
"rewards/margins": 3.946502685546875,
|
| 1054 |
+
"rewards/rejected": -11.193906784057617,
|
| 1055 |
+
"step": 580
|
| 1056 |
+
},
|
| 1057 |
+
{
|
| 1058 |
+
"epoch": 0.9260349225034334,
|
| 1059 |
+
"grad_norm": 3.3875763416290283,
|
| 1060 |
+
"learning_rate": 3.3221666168464584e-06,
|
| 1061 |
+
"logits/chosen": 12.595063209533691,
|
| 1062 |
+
"logits/rejected": 12.006935119628906,
|
| 1063 |
+
"logps/chosen": -4.957076072692871,
|
| 1064 |
+
"logps/rejected": -7.576946258544922,
|
| 1065 |
+
"loss": 0.3553,
|
| 1066 |
+
"rewards/accuracies": 0.887499988079071,
|
| 1067 |
+
"rewards/chosen": -7.435614109039307,
|
| 1068 |
+
"rewards/margins": 3.9298064708709717,
|
| 1069 |
+
"rewards/rejected": -11.3654203414917,
|
| 1070 |
+
"step": 590
|
| 1071 |
+
},
|
| 1072 |
+
{
|
| 1073 |
+
"epoch": 0.9417304296645085,
|
| 1074 |
+
"grad_norm": 2.784954071044922,
|
| 1075 |
+
"learning_rate": 3.272542485937369e-06,
|
| 1076 |
+
"logits/chosen": 12.631284713745117,
|
| 1077 |
+
"logits/rejected": 11.908422470092773,
|
| 1078 |
+
"logps/chosen": -4.122859477996826,
|
| 1079 |
+
"logps/rejected": -7.009615421295166,
|
| 1080 |
+
"loss": 0.3086,
|
| 1081 |
+
"rewards/accuracies": 0.875,
|
| 1082 |
+
"rewards/chosen": -6.18428897857666,
|
| 1083 |
+
"rewards/margins": 4.330134868621826,
|
| 1084 |
+
"rewards/rejected": -10.514424324035645,
|
| 1085 |
+
"step": 600
|
| 1086 |
+
},
|
| 1087 |
+
{
|
| 1088 |
+
"epoch": 0.9417304296645085,
|
| 1089 |
+
"eval_logits/chosen": 13.370320320129395,
|
| 1090 |
+
"eval_logits/rejected": 12.268073081970215,
|
| 1091 |
+
"eval_logps/chosen": -4.0377197265625,
|
| 1092 |
+
"eval_logps/rejected": -7.431212425231934,
|
| 1093 |
+
"eval_loss": 0.33698391914367676,
|
| 1094 |
+
"eval_rewards/accuracies": 1.0,
|
| 1095 |
+
"eval_rewards/chosen": -6.056579113006592,
|
| 1096 |
+
"eval_rewards/margins": 5.090238094329834,
|
| 1097 |
+
"eval_rewards/rejected": -11.146818161010742,
|
| 1098 |
+
"eval_runtime": 14.5012,
|
| 1099 |
+
"eval_samples_per_second": 28.411,
|
| 1100 |
+
"eval_steps_per_second": 3.586,
|
| 1101 |
+
"step": 600
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"epoch": 0.9574259368255836,
|
| 1105 |
+
"grad_norm": 3.2864303588867188,
|
| 1106 |
+
"learning_rate": 3.222579492361179e-06,
|
| 1107 |
+
"logits/chosen": 12.638185501098633,
|
| 1108 |
+
"logits/rejected": 11.773119926452637,
|
| 1109 |
+
"logps/chosen": -4.101125240325928,
|
| 1110 |
+
"logps/rejected": -6.985491752624512,
|
| 1111 |
+
"loss": 0.2886,
|
| 1112 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 1113 |
+
"rewards/chosen": -6.1516876220703125,
|
| 1114 |
+
"rewards/margins": 4.326549530029297,
|
| 1115 |
+
"rewards/rejected": -10.47823715209961,
|
| 1116 |
+
"step": 610
|
| 1117 |
+
},
|
| 1118 |
+
{
|
| 1119 |
+
"epoch": 0.9731214439866588,
|
| 1120 |
+
"grad_norm": 3.094020366668701,
|
| 1121 |
+
"learning_rate": 3.1722995515381644e-06,
|
| 1122 |
+
"logits/chosen": 12.71407413482666,
|
| 1123 |
+
"logits/rejected": 11.668571472167969,
|
| 1124 |
+
"logps/chosen": -4.560760498046875,
|
| 1125 |
+
"logps/rejected": -7.878456115722656,
|
| 1126 |
+
"loss": 0.294,
|
| 1127 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 1128 |
+
"rewards/chosen": -6.841139793395996,
|
| 1129 |
+
"rewards/margins": 4.976544380187988,
|
| 1130 |
+
"rewards/rejected": -11.817686080932617,
|
| 1131 |
+
"step": 620
|
| 1132 |
+
},
|
| 1133 |
+
{
|
| 1134 |
+
"epoch": 0.9888169511477339,
|
| 1135 |
+
"grad_norm": 1.906849980354309,
|
| 1136 |
+
"learning_rate": 3.121724717912138e-06,
|
| 1137 |
+
"logits/chosen": 12.863229751586914,
|
| 1138 |
+
"logits/rejected": 11.391074180603027,
|
| 1139 |
+
"logps/chosen": -4.578230857849121,
|
| 1140 |
+
"logps/rejected": -8.216089248657227,
|
| 1141 |
+
"loss": 0.2579,
|
| 1142 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 1143 |
+
"rewards/chosen": -6.86734676361084,
|
| 1144 |
+
"rewards/margins": 5.456785202026367,
|
| 1145 |
+
"rewards/rejected": -12.324131965637207,
|
| 1146 |
+
"step": 630
|
| 1147 |
+
},
|
| 1148 |
+
{
|
| 1149 |
+
"epoch": 1.003139101432215,
|
| 1150 |
+
"grad_norm": 2.699907064437866,
|
| 1151 |
+
"learning_rate": 3.0708771752766397e-06,
|
| 1152 |
+
"logits/chosen": 12.64507007598877,
|
| 1153 |
+
"logits/rejected": 12.148189544677734,
|
| 1154 |
+
"logps/chosen": -4.193073272705078,
|
| 1155 |
+
"logps/rejected": -7.2082905769348145,
|
| 1156 |
+
"loss": 0.2449,
|
| 1157 |
+
"rewards/accuracies": 0.931506872177124,
|
| 1158 |
+
"rewards/chosen": -6.289609909057617,
|
| 1159 |
+
"rewards/margins": 4.522825241088867,
|
| 1160 |
+
"rewards/rejected": -10.812435150146484,
|
| 1161 |
+
"step": 640
|
| 1162 |
+
},
|
| 1163 |
+
{
|
| 1164 |
+
"epoch": 1.0188346085932902,
|
| 1165 |
+
"grad_norm": 3.4014647006988525,
|
| 1166 |
+
"learning_rate": 3.019779227044398e-06,
|
| 1167 |
+
"logits/chosen": 12.921002388000488,
|
| 1168 |
+
"logits/rejected": 12.00908088684082,
|
| 1169 |
+
"logps/chosen": -4.350411415100098,
|
| 1170 |
+
"logps/rejected": -7.187474250793457,
|
| 1171 |
+
"loss": 0.2976,
|
| 1172 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 1173 |
+
"rewards/chosen": -6.5256171226501465,
|
| 1174 |
+
"rewards/margins": 4.255594253540039,
|
| 1175 |
+
"rewards/rejected": -10.781211853027344,
|
| 1176 |
+
"step": 650
|
| 1177 |
+
},
|
| 1178 |
+
{
|
| 1179 |
+
"epoch": 1.0188346085932902,
|
| 1180 |
+
"eval_logits/chosen": 13.153624534606934,
|
| 1181 |
+
"eval_logits/rejected": 12.077543258666992,
|
| 1182 |
+
"eval_logps/chosen": -4.4063401222229,
|
| 1183 |
+
"eval_logps/rejected": -8.090719223022461,
|
| 1184 |
+
"eval_loss": 0.31872284412384033,
|
| 1185 |
+
"eval_rewards/accuracies": 1.0,
|
| 1186 |
+
"eval_rewards/chosen": -6.6095099449157715,
|
| 1187 |
+
"eval_rewards/margins": 5.526568412780762,
|
| 1188 |
+
"eval_rewards/rejected": -12.136077880859375,
|
| 1189 |
+
"eval_runtime": 14.4992,
|
| 1190 |
+
"eval_samples_per_second": 28.415,
|
| 1191 |
+
"eval_steps_per_second": 3.586,
|
| 1192 |
+
"step": 650
|
| 1193 |
+
},
|
| 1194 |
+
{
|
| 1195 |
+
"epoch": 1.0345301157543654,
|
| 1196 |
+
"grad_norm": 3.4320549964904785,
|
| 1197 |
+
"learning_rate": 2.9684532864643123e-06,
|
| 1198 |
+
"logits/chosen": 12.755729675292969,
|
| 1199 |
+
"logits/rejected": 11.981709480285645,
|
| 1200 |
+
"logps/chosen": -4.497554302215576,
|
| 1201 |
+
"logps/rejected": -7.250922203063965,
|
| 1202 |
+
"loss": 0.3013,
|
| 1203 |
+
"rewards/accuracies": 0.887499988079071,
|
| 1204 |
+
"rewards/chosen": -6.746331691741943,
|
| 1205 |
+
"rewards/margins": 4.13005256652832,
|
| 1206 |
+
"rewards/rejected": -10.876383781433105,
|
| 1207 |
+
"step": 660
|
| 1208 |
+
},
|
| 1209 |
+
{
|
| 1210 |
+
"epoch": 1.0502256229154405,
|
| 1211 |
+
"grad_norm": 5.130344867706299,
|
| 1212 |
+
"learning_rate": 2.9169218667902562e-06,
|
| 1213 |
+
"logits/chosen": 12.774751663208008,
|
| 1214 |
+
"logits/rejected": 12.063162803649902,
|
| 1215 |
+
"logps/chosen": -4.672727108001709,
|
| 1216 |
+
"logps/rejected": -7.972691535949707,
|
| 1217 |
+
"loss": 0.2677,
|
| 1218 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 1219 |
+
"rewards/chosen": -7.009090423583984,
|
| 1220 |
+
"rewards/margins": 4.94994592666626,
|
| 1221 |
+
"rewards/rejected": -11.959035873413086,
|
| 1222 |
+
"step": 670
|
| 1223 |
+
},
|
| 1224 |
+
{
|
| 1225 |
+
"epoch": 1.0659211300765157,
|
| 1226 |
+
"grad_norm": 2.1107585430145264,
|
| 1227 |
+
"learning_rate": 2.8652075714060296e-06,
|
| 1228 |
+
"logits/chosen": 12.401437759399414,
|
| 1229 |
+
"logits/rejected": 11.329314231872559,
|
| 1230 |
+
"logps/chosen": -4.372309684753418,
|
| 1231 |
+
"logps/rejected": -7.968375205993652,
|
| 1232 |
+
"loss": 0.2321,
|
| 1233 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1234 |
+
"rewards/chosen": -6.558464050292969,
|
| 1235 |
+
"rewards/margins": 5.394099235534668,
|
| 1236 |
+
"rewards/rejected": -11.952563285827637,
|
| 1237 |
+
"step": 680
|
| 1238 |
+
},
|
| 1239 |
+
{
|
| 1240 |
+
"epoch": 1.0816166372375908,
|
| 1241 |
+
"grad_norm": 2.09533429145813,
|
| 1242 |
+
"learning_rate": 2.813333083910761e-06,
|
| 1243 |
+
"logits/chosen": 12.615015029907227,
|
| 1244 |
+
"logits/rejected": 11.613394737243652,
|
| 1245 |
+
"logps/chosen": -4.582761764526367,
|
| 1246 |
+
"logps/rejected": -7.808516502380371,
|
| 1247 |
+
"loss": 0.2593,
|
| 1248 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 1249 |
+
"rewards/chosen": -6.874144077301025,
|
| 1250 |
+
"rewards/margins": 4.838631629943848,
|
| 1251 |
+
"rewards/rejected": -11.712774276733398,
|
| 1252 |
+
"step": 690
|
| 1253 |
+
},
|
| 1254 |
+
{
|
| 1255 |
+
"epoch": 1.097312144398666,
|
| 1256 |
+
"grad_norm": 2.652644157409668,
|
| 1257 |
+
"learning_rate": 2.761321158169134e-06,
|
| 1258 |
+
"logits/chosen": 12.447229385375977,
|
| 1259 |
+
"logits/rejected": 11.710171699523926,
|
| 1260 |
+
"logps/chosen": -4.722737789154053,
|
| 1261 |
+
"logps/rejected": -8.122079849243164,
|
| 1262 |
+
"loss": 0.263,
|
| 1263 |
+
"rewards/accuracies": 0.9375,
|
| 1264 |
+
"rewards/chosen": -7.084107398986816,
|
| 1265 |
+
"rewards/margins": 5.0990142822265625,
|
| 1266 |
+
"rewards/rejected": -12.183120727539062,
|
| 1267 |
+
"step": 700
|
| 1268 |
+
},
|
| 1269 |
+
{
|
| 1270 |
+
"epoch": 1.097312144398666,
|
| 1271 |
+
"eval_logits/chosen": 13.170889854431152,
|
| 1272 |
+
"eval_logits/rejected": 12.095331192016602,
|
| 1273 |
+
"eval_logps/chosen": -4.405147075653076,
|
| 1274 |
+
"eval_logps/rejected": -8.16016960144043,
|
| 1275 |
+
"eval_loss": 0.3023497760295868,
|
| 1276 |
+
"eval_rewards/accuracies": 1.0,
|
| 1277 |
+
"eval_rewards/chosen": -6.607721328735352,
|
| 1278 |
+
"eval_rewards/margins": 5.632532119750977,
|
| 1279 |
+
"eval_rewards/rejected": -12.240253448486328,
|
| 1280 |
+
"eval_runtime": 14.5027,
|
| 1281 |
+
"eval_samples_per_second": 28.408,
|
| 1282 |
+
"eval_steps_per_second": 3.586,
|
| 1283 |
+
"step": 700
|
| 1284 |
+
},
|
| 1285 |
+
{
|
| 1286 |
+
"epoch": 1.113007651559741,
|
| 1287 |
+
"grad_norm": 2.3017961978912354,
|
| 1288 |
+
"learning_rate": 2.70919460833079e-06,
|
| 1289 |
+
"logits/chosen": 12.642290115356445,
|
| 1290 |
+
"logits/rejected": 11.817375183105469,
|
| 1291 |
+
"logps/chosen": -4.732531547546387,
|
| 1292 |
+
"logps/rejected": -8.388056755065918,
|
| 1293 |
+
"loss": 0.2707,
|
| 1294 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1295 |
+
"rewards/chosen": -7.098797798156738,
|
| 1296 |
+
"rewards/margins": 5.483288764953613,
|
| 1297 |
+
"rewards/rejected": -12.582086563110352,
|
| 1298 |
+
"step": 710
|
| 1299 |
+
},
|
| 1300 |
+
{
|
| 1301 |
+
"epoch": 1.1287031587208163,
|
| 1302 |
+
"grad_norm": 2.538917303085327,
|
| 1303 |
+
"learning_rate": 2.6569762988232838e-06,
|
| 1304 |
+
"logits/chosen": 12.873556137084961,
|
| 1305 |
+
"logits/rejected": 11.74571704864502,
|
| 1306 |
+
"logps/chosen": -5.0908284187316895,
|
| 1307 |
+
"logps/rejected": -8.69284439086914,
|
| 1308 |
+
"loss": 0.2443,
|
| 1309 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 1310 |
+
"rewards/chosen": -7.636242866516113,
|
| 1311 |
+
"rewards/margins": 5.403023719787598,
|
| 1312 |
+
"rewards/rejected": -13.039266586303711,
|
| 1313 |
+
"step": 720
|
| 1314 |
+
},
|
| 1315 |
+
{
|
| 1316 |
+
"epoch": 1.1443986658818912,
|
| 1317 |
+
"grad_norm": 2.7106313705444336,
|
| 1318 |
+
"learning_rate": 2.604689134322999e-06,
|
| 1319 |
+
"logits/chosen": 12.663686752319336,
|
| 1320 |
+
"logits/rejected": 12.057671546936035,
|
| 1321 |
+
"logps/chosen": -4.6052398681640625,
|
| 1322 |
+
"logps/rejected": -7.64666748046875,
|
| 1323 |
+
"loss": 0.2657,
|
| 1324 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 1325 |
+
"rewards/chosen": -6.907858848571777,
|
| 1326 |
+
"rewards/margins": 4.562141418457031,
|
| 1327 |
+
"rewards/rejected": -11.470001220703125,
|
| 1328 |
+
"step": 730
|
| 1329 |
+
},
|
| 1330 |
+
{
|
| 1331 |
+
"epoch": 1.1600941730429664,
|
| 1332 |
+
"grad_norm": 1.9821428060531616,
|
| 1333 |
+
"learning_rate": 2.5523560497083927e-06,
|
| 1334 |
+
"logits/chosen": 12.605002403259277,
|
| 1335 |
+
"logits/rejected": 11.786115646362305,
|
| 1336 |
+
"logps/chosen": -4.447022438049316,
|
| 1337 |
+
"logps/rejected": -8.232423782348633,
|
| 1338 |
+
"loss": 0.2234,
|
| 1339 |
+
"rewards/accuracies": 0.9375,
|
| 1340 |
+
"rewards/chosen": -6.670533180236816,
|
| 1341 |
+
"rewards/margins": 5.678102016448975,
|
| 1342 |
+
"rewards/rejected": -12.348634719848633,
|
| 1343 |
+
"step": 740
|
| 1344 |
+
},
|
| 1345 |
+
{
|
| 1346 |
+
"epoch": 1.1757896802040415,
|
| 1347 |
+
"grad_norm": 2.469222068786621,
|
| 1348 |
+
"learning_rate": 2.5e-06,
|
| 1349 |
+
"logits/chosen": 12.427464485168457,
|
| 1350 |
+
"logits/rejected": 11.602205276489258,
|
| 1351 |
+
"logps/chosen": -4.620213985443115,
|
| 1352 |
+
"logps/rejected": -8.116684913635254,
|
| 1353 |
+
"loss": 0.2895,
|
| 1354 |
+
"rewards/accuracies": 0.887499988079071,
|
| 1355 |
+
"rewards/chosen": -6.93032169342041,
|
| 1356 |
+
"rewards/margins": 5.244706153869629,
|
| 1357 |
+
"rewards/rejected": -12.175026893615723,
|
| 1358 |
+
"step": 750
|
| 1359 |
+
},
|
| 1360 |
+
{
|
| 1361 |
+
"epoch": 1.1757896802040415,
|
| 1362 |
+
"eval_logits/chosen": 13.117497444152832,
|
| 1363 |
+
"eval_logits/rejected": 12.068766593933105,
|
| 1364 |
+
"eval_logps/chosen": -4.3247389793396,
|
| 1365 |
+
"eval_logps/rejected": -8.284286499023438,
|
| 1366 |
+
"eval_loss": 0.2885077893733978,
|
| 1367 |
+
"eval_rewards/accuracies": 1.0,
|
| 1368 |
+
"eval_rewards/chosen": -6.48710823059082,
|
| 1369 |
+
"eval_rewards/margins": 5.939322471618652,
|
| 1370 |
+
"eval_rewards/rejected": -12.426431655883789,
|
| 1371 |
+
"eval_runtime": 14.5056,
|
| 1372 |
+
"eval_samples_per_second": 28.403,
|
| 1373 |
+
"eval_steps_per_second": 3.585,
|
| 1374 |
+
"step": 750
|
| 1375 |
+
},
|
| 1376 |
+
{
|
| 1377 |
+
"epoch": 1.1914851873651167,
|
| 1378 |
+
"grad_norm": 1.4759465456008911,
|
| 1379 |
+
"learning_rate": 2.447643950291608e-06,
|
| 1380 |
+
"logits/chosen": 12.716300010681152,
|
| 1381 |
+
"logits/rejected": 11.687711715698242,
|
| 1382 |
+
"logps/chosen": -4.740975379943848,
|
| 1383 |
+
"logps/rejected": -8.425516128540039,
|
| 1384 |
+
"loss": 0.2696,
|
| 1385 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 1386 |
+
"rewards/chosen": -7.111462593078613,
|
| 1387 |
+
"rewards/margins": 5.526811122894287,
|
| 1388 |
+
"rewards/rejected": -12.638272285461426,
|
| 1389 |
+
"step": 760
|
| 1390 |
+
},
|
| 1391 |
+
{
|
| 1392 |
+
"epoch": 1.2071806945261918,
|
| 1393 |
+
"grad_norm": 4.352353572845459,
|
| 1394 |
+
"learning_rate": 2.3953108656770018e-06,
|
| 1395 |
+
"logits/chosen": 12.28320598602295,
|
| 1396 |
+
"logits/rejected": 11.589045524597168,
|
| 1397 |
+
"logps/chosen": -4.623471260070801,
|
| 1398 |
+
"logps/rejected": -7.964238166809082,
|
| 1399 |
+
"loss": 0.2692,
|
| 1400 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 1401 |
+
"rewards/chosen": -6.935206413269043,
|
| 1402 |
+
"rewards/margins": 5.01115083694458,
|
| 1403 |
+
"rewards/rejected": -11.946358680725098,
|
| 1404 |
+
"step": 770
|
| 1405 |
+
},
|
| 1406 |
+
{
|
| 1407 |
+
"epoch": 1.222876201687267,
|
| 1408 |
+
"grad_norm": 3.407118558883667,
|
| 1409 |
+
"learning_rate": 2.3430237011767166e-06,
|
| 1410 |
+
"logits/chosen": 12.472453117370605,
|
| 1411 |
+
"logits/rejected": 11.568643569946289,
|
| 1412 |
+
"logps/chosen": -4.998484134674072,
|
| 1413 |
+
"logps/rejected": -9.068008422851562,
|
| 1414 |
+
"loss": 0.248,
|
| 1415 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 1416 |
+
"rewards/chosen": -7.4977264404296875,
|
| 1417 |
+
"rewards/margins": 6.104286193847656,
|
| 1418 |
+
"rewards/rejected": -13.602012634277344,
|
| 1419 |
+
"step": 780
|
| 1420 |
+
},
|
| 1421 |
+
{
|
| 1422 |
+
"epoch": 1.2385717088483421,
|
| 1423 |
+
"grad_norm": 2.2443063259124756,
|
| 1424 |
+
"learning_rate": 2.290805391669212e-06,
|
| 1425 |
+
"logits/chosen": 12.315655708312988,
|
| 1426 |
+
"logits/rejected": 11.502891540527344,
|
| 1427 |
+
"logps/chosen": -4.696177959442139,
|
| 1428 |
+
"logps/rejected": -8.123588562011719,
|
| 1429 |
+
"loss": 0.2549,
|
| 1430 |
+
"rewards/accuracies": 1.0,
|
| 1431 |
+
"rewards/chosen": -7.044266700744629,
|
| 1432 |
+
"rewards/margins": 5.141116142272949,
|
| 1433 |
+
"rewards/rejected": -12.185382843017578,
|
| 1434 |
+
"step": 790
|
| 1435 |
+
},
|
| 1436 |
+
{
|
| 1437 |
+
"epoch": 1.2542672160094173,
|
| 1438 |
+
"grad_norm": 3.419119358062744,
|
| 1439 |
+
"learning_rate": 2.238678841830867e-06,
|
| 1440 |
+
"logits/chosen": 12.070470809936523,
|
| 1441 |
+
"logits/rejected": 11.558723449707031,
|
| 1442 |
+
"logps/chosen": -4.473680019378662,
|
| 1443 |
+
"logps/rejected": -7.936869144439697,
|
| 1444 |
+
"loss": 0.3082,
|
| 1445 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1446 |
+
"rewards/chosen": -6.710520267486572,
|
| 1447 |
+
"rewards/margins": 5.1947832107543945,
|
| 1448 |
+
"rewards/rejected": -11.905302047729492,
|
| 1449 |
+
"step": 800
|
| 1450 |
+
},
|
| 1451 |
+
{
|
| 1452 |
+
"epoch": 1.2542672160094173,
|
| 1453 |
+
"eval_logits/chosen": 13.084930419921875,
|
| 1454 |
+
"eval_logits/rejected": 12.024311065673828,
|
| 1455 |
+
"eval_logps/chosen": -4.1232075691223145,
|
| 1456 |
+
"eval_logps/rejected": -8.072820663452148,
|
| 1457 |
+
"eval_loss": 0.27581357955932617,
|
| 1458 |
+
"eval_rewards/accuracies": 1.0,
|
| 1459 |
+
"eval_rewards/chosen": -6.184810161590576,
|
| 1460 |
+
"eval_rewards/margins": 5.924418926239014,
|
| 1461 |
+
"eval_rewards/rejected": -12.10922908782959,
|
| 1462 |
+
"eval_runtime": 14.5062,
|
| 1463 |
+
"eval_samples_per_second": 28.402,
|
| 1464 |
+
"eval_steps_per_second": 3.585,
|
| 1465 |
+
"step": 800
|
| 1466 |
+
},
|
| 1467 |
+
{
|
| 1468 |
+
"epoch": 1.2699627231704924,
|
| 1469 |
+
"grad_norm": 3.637188196182251,
|
| 1470 |
+
"learning_rate": 2.186666916089239e-06,
|
| 1471 |
+
"logits/chosen": 12.425226211547852,
|
| 1472 |
+
"logits/rejected": 11.625535011291504,
|
| 1473 |
+
"logps/chosen": -4.465217590332031,
|
| 1474 |
+
"logps/rejected": -8.138174057006836,
|
| 1475 |
+
"loss": 0.2454,
|
| 1476 |
+
"rewards/accuracies": 0.9375,
|
| 1477 |
+
"rewards/chosen": -6.697826385498047,
|
| 1478 |
+
"rewards/margins": 5.509435653686523,
|
| 1479 |
+
"rewards/rejected": -12.20726203918457,
|
| 1480 |
+
"step": 810
|
| 1481 |
+
},
|
| 1482 |
+
{
|
| 1483 |
+
"epoch": 1.2856582303315676,
|
| 1484 |
+
"grad_norm": 2.836642265319824,
|
| 1485 |
+
"learning_rate": 2.134792428593971e-06,
|
| 1486 |
+
"logits/chosen": 12.305977821350098,
|
| 1487 |
+
"logits/rejected": 11.66655445098877,
|
| 1488 |
+
"logps/chosen": -4.680156707763672,
|
| 1489 |
+
"logps/rejected": -7.750007629394531,
|
| 1490 |
+
"loss": 0.2172,
|
| 1491 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 1492 |
+
"rewards/chosen": -7.020236015319824,
|
| 1493 |
+
"rewards/margins": 4.604775905609131,
|
| 1494 |
+
"rewards/rejected": -11.625011444091797,
|
| 1495 |
+
"step": 820
|
| 1496 |
+
},
|
| 1497 |
+
{
|
| 1498 |
+
"epoch": 1.3013537374926427,
|
| 1499 |
+
"grad_norm": 3.5275964736938477,
|
| 1500 |
+
"learning_rate": 2.0830781332097446e-06,
|
| 1501 |
+
"logits/chosen": 12.902934074401855,
|
| 1502 |
+
"logits/rejected": 11.944432258605957,
|
| 1503 |
+
"logps/chosen": -4.6034321784973145,
|
| 1504 |
+
"logps/rejected": -8.786798477172852,
|
| 1505 |
+
"loss": 0.2231,
|
| 1506 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1507 |
+
"rewards/chosen": -6.905147552490234,
|
| 1508 |
+
"rewards/margins": 6.275050163269043,
|
| 1509 |
+
"rewards/rejected": -13.180196762084961,
|
| 1510 |
+
"step": 830
|
| 1511 |
+
},
|
| 1512 |
+
{
|
| 1513 |
+
"epoch": 1.3170492446537179,
|
| 1514 |
+
"grad_norm": 3.2921273708343506,
|
| 1515 |
+
"learning_rate": 2.031546713535688e-06,
|
| 1516 |
+
"logits/chosen": 12.8768892288208,
|
| 1517 |
+
"logits/rejected": 11.985623359680176,
|
| 1518 |
+
"logps/chosen": -4.825736045837402,
|
| 1519 |
+
"logps/rejected": -8.161986351013184,
|
| 1520 |
+
"loss": 0.2584,
|
| 1521 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 1522 |
+
"rewards/chosen": -7.2386040687561035,
|
| 1523 |
+
"rewards/margins": 5.00437593460083,
|
| 1524 |
+
"rewards/rejected": -12.242980003356934,
|
| 1525 |
+
"step": 840
|
| 1526 |
+
},
|
| 1527 |
+
{
|
| 1528 |
+
"epoch": 1.332744751814793,
|
| 1529 |
+
"grad_norm": 3.1977927684783936,
|
| 1530 |
+
"learning_rate": 1.9802207729556023e-06,
|
| 1531 |
+
"logits/chosen": 12.695280075073242,
|
| 1532 |
+
"logits/rejected": 11.713974952697754,
|
| 1533 |
+
"logps/chosen": -4.446492671966553,
|
| 1534 |
+
"logps/rejected": -8.236989974975586,
|
| 1535 |
+
"loss": 0.2242,
|
| 1536 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 1537 |
+
"rewards/chosen": -6.669739723205566,
|
| 1538 |
+
"rewards/margins": 5.685745716094971,
|
| 1539 |
+
"rewards/rejected": -12.355484008789062,
|
| 1540 |
+
"step": 850
|
| 1541 |
+
},
|
| 1542 |
+
{
|
| 1543 |
+
"epoch": 1.332744751814793,
|
| 1544 |
+
"eval_logits/chosen": 12.975180625915527,
|
| 1545 |
+
"eval_logits/rejected": 11.917383193969727,
|
| 1546 |
+
"eval_logps/chosen": -4.095187664031982,
|
| 1547 |
+
"eval_logps/rejected": -8.178255081176758,
|
| 1548 |
+
"eval_loss": 0.2621641755104065,
|
| 1549 |
+
"eval_rewards/accuracies": 1.0,
|
| 1550 |
+
"eval_rewards/chosen": -6.1427812576293945,
|
| 1551 |
+
"eval_rewards/margins": 6.124600887298584,
|
| 1552 |
+
"eval_rewards/rejected": -12.26738166809082,
|
| 1553 |
+
"eval_runtime": 14.5043,
|
| 1554 |
+
"eval_samples_per_second": 28.405,
|
| 1555 |
+
"eval_steps_per_second": 3.585,
|
| 1556 |
+
"step": 850
|
| 1557 |
+
},
|
| 1558 |
+
{
|
| 1559 |
+
"epoch": 1.3484402589758682,
|
| 1560 |
+
"grad_norm": 4.474134922027588,
|
| 1561 |
+
"learning_rate": 1.9291228247233607e-06,
|
| 1562 |
+
"logits/chosen": 12.378435134887695,
|
| 1563 |
+
"logits/rejected": 11.085493087768555,
|
| 1564 |
+
"logps/chosen": -4.5196709632873535,
|
| 1565 |
+
"logps/rejected": -8.377789497375488,
|
| 1566 |
+
"loss": 0.2711,
|
| 1567 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 1568 |
+
"rewards/chosen": -6.779506683349609,
|
| 1569 |
+
"rewards/margins": 5.787177085876465,
|
| 1570 |
+
"rewards/rejected": -12.566683769226074,
|
| 1571 |
+
"step": 860
|
| 1572 |
+
},
|
| 1573 |
+
{
|
| 1574 |
+
"epoch": 1.3641357661369433,
|
| 1575 |
+
"grad_norm": 3.2606918811798096,
|
| 1576 |
+
"learning_rate": 1.8782752820878636e-06,
|
| 1577 |
+
"logits/chosen": 11.961488723754883,
|
| 1578 |
+
"logits/rejected": 11.434579849243164,
|
| 1579 |
+
"logps/chosen": -4.525224208831787,
|
| 1580 |
+
"logps/rejected": -8.16456413269043,
|
| 1581 |
+
"loss": 0.2569,
|
| 1582 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 1583 |
+
"rewards/chosen": -6.787835597991943,
|
| 1584 |
+
"rewards/margins": 5.459010124206543,
|
| 1585 |
+
"rewards/rejected": -12.246846199035645,
|
| 1586 |
+
"step": 870
|
| 1587 |
+
},
|
| 1588 |
+
{
|
| 1589 |
+
"epoch": 1.3798312732980185,
|
| 1590 |
+
"grad_norm": 3.297881603240967,
|
| 1591 |
+
"learning_rate": 1.827700448461836e-06,
|
| 1592 |
+
"logits/chosen": 12.268153190612793,
|
| 1593 |
+
"logits/rejected": 11.848055839538574,
|
| 1594 |
+
"logps/chosen": -3.979138135910034,
|
| 1595 |
+
"logps/rejected": -7.244141578674316,
|
| 1596 |
+
"loss": 0.2876,
|
| 1597 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1598 |
+
"rewards/chosen": -5.968707084655762,
|
| 1599 |
+
"rewards/margins": 4.897505760192871,
|
| 1600 |
+
"rewards/rejected": -10.866212844848633,
|
| 1601 |
+
"step": 880
|
| 1602 |
+
},
|
| 1603 |
+
{
|
| 1604 |
+
"epoch": 1.3955267804590936,
|
| 1605 |
+
"grad_norm": 5.355867862701416,
|
| 1606 |
+
"learning_rate": 1.7774205076388207e-06,
|
| 1607 |
+
"logits/chosen": 12.47008991241455,
|
| 1608 |
+
"logits/rejected": 11.850513458251953,
|
| 1609 |
+
"logps/chosen": -4.1252217292785645,
|
| 1610 |
+
"logps/rejected": -8.060781478881836,
|
| 1611 |
+
"loss": 0.2502,
|
| 1612 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1613 |
+
"rewards/chosen": -6.187832832336426,
|
| 1614 |
+
"rewards/margins": 5.9033403396606445,
|
| 1615 |
+
"rewards/rejected": -12.09117317199707,
|
| 1616 |
+
"step": 890
|
| 1617 |
+
},
|
| 1618 |
+
{
|
| 1619 |
+
"epoch": 1.4112222876201688,
|
| 1620 |
+
"grad_norm": 3.333230495452881,
|
| 1621 |
+
"learning_rate": 1.7274575140626318e-06,
|
| 1622 |
+
"logits/chosen": 12.16104793548584,
|
| 1623 |
+
"logits/rejected": 11.287874221801758,
|
| 1624 |
+
"logps/chosen": -3.827984571456909,
|
| 1625 |
+
"logps/rejected": -7.516768455505371,
|
| 1626 |
+
"loss": 0.2162,
|
| 1627 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 1628 |
+
"rewards/chosen": -5.741976737976074,
|
| 1629 |
+
"rewards/margins": 5.533174991607666,
|
| 1630 |
+
"rewards/rejected": -11.275151252746582,
|
| 1631 |
+
"step": 900
|
| 1632 |
+
},
|
| 1633 |
+
{
|
| 1634 |
+
"epoch": 1.4112222876201688,
|
| 1635 |
+
"eval_logits/chosen": 12.950577735900879,
|
| 1636 |
+
"eval_logits/rejected": 11.89844799041748,
|
| 1637 |
+
"eval_logps/chosen": -4.105511665344238,
|
| 1638 |
+
"eval_logps/rejected": -8.293257713317871,
|
| 1639 |
+
"eval_loss": 0.2529076933860779,
|
| 1640 |
+
"eval_rewards/accuracies": 1.0,
|
| 1641 |
+
"eval_rewards/chosen": -6.158267498016357,
|
| 1642 |
+
"eval_rewards/margins": 6.281619071960449,
|
| 1643 |
+
"eval_rewards/rejected": -12.439887046813965,
|
| 1644 |
+
"eval_runtime": 14.5023,
|
| 1645 |
+
"eval_samples_per_second": 28.409,
|
| 1646 |
+
"eval_steps_per_second": 3.586,
|
| 1647 |
+
"step": 900
|
| 1648 |
+
},
|
| 1649 |
+
{
|
| 1650 |
+
"epoch": 1.426917794781244,
|
| 1651 |
+
"grad_norm": 2.3288955688476562,
|
| 1652 |
+
"learning_rate": 1.677833383153542e-06,
|
| 1653 |
+
"logits/chosen": 12.872639656066895,
|
| 1654 |
+
"logits/rejected": 11.605780601501465,
|
| 1655 |
+
"logps/chosen": -4.383180141448975,
|
| 1656 |
+
"logps/rejected": -8.010499954223633,
|
| 1657 |
+
"loss": 0.212,
|
| 1658 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1659 |
+
"rewards/chosen": -6.574770927429199,
|
| 1660 |
+
"rewards/margins": 5.440978527069092,
|
| 1661 |
+
"rewards/rejected": -12.015748977661133,
|
| 1662 |
+
"step": 910
|
| 1663 |
+
},
|
| 1664 |
+
{
|
| 1665 |
+
"epoch": 1.442613301942319,
|
| 1666 |
+
"grad_norm": 3.773186206817627,
|
| 1667 |
+
"learning_rate": 1.6285698816954626e-06,
|
| 1668 |
+
"logits/chosen": 12.562664985656738,
|
| 1669 |
+
"logits/rejected": 11.542292594909668,
|
| 1670 |
+
"logps/chosen": -4.957005500793457,
|
| 1671 |
+
"logps/rejected": -9.080846786499023,
|
| 1672 |
+
"loss": 0.2343,
|
| 1673 |
+
"rewards/accuracies": 0.9375,
|
| 1674 |
+
"rewards/chosen": -7.435508728027344,
|
| 1675 |
+
"rewards/margins": 6.185763359069824,
|
| 1676 |
+
"rewards/rejected": -13.621271133422852,
|
| 1677 |
+
"step": 920
|
| 1678 |
+
},
|
| 1679 |
+
{
|
| 1680 |
+
"epoch": 1.4583088091033942,
|
| 1681 |
+
"grad_norm": 2.7567288875579834,
|
| 1682 |
+
"learning_rate": 1.5796886182883053e-06,
|
| 1683 |
+
"logits/chosen": 12.21378231048584,
|
| 1684 |
+
"logits/rejected": 11.380667686462402,
|
| 1685 |
+
"logps/chosen": -4.714855670928955,
|
| 1686 |
+
"logps/rejected": -8.239659309387207,
|
| 1687 |
+
"loss": 0.2487,
|
| 1688 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1689 |
+
"rewards/chosen": -7.072283744812012,
|
| 1690 |
+
"rewards/margins": 5.287206172943115,
|
| 1691 |
+
"rewards/rejected": -12.359490394592285,
|
| 1692 |
+
"step": 930
|
| 1693 |
+
},
|
| 1694 |
+
{
|
| 1695 |
+
"epoch": 1.4740043162644694,
|
| 1696 |
+
"grad_norm": 5.543617248535156,
|
| 1697 |
+
"learning_rate": 1.5312110338697427e-06,
|
| 1698 |
+
"logits/chosen": 11.844781875610352,
|
| 1699 |
+
"logits/rejected": 11.141448974609375,
|
| 1700 |
+
"logps/chosen": -4.98689079284668,
|
| 1701 |
+
"logps/rejected": -8.525138854980469,
|
| 1702 |
+
"loss": 0.2609,
|
| 1703 |
+
"rewards/accuracies": 0.9375,
|
| 1704 |
+
"rewards/chosen": -7.480336666107178,
|
| 1705 |
+
"rewards/margins": 5.307372570037842,
|
| 1706 |
+
"rewards/rejected": -12.78770923614502,
|
| 1707 |
+
"step": 940
|
| 1708 |
+
},
|
| 1709 |
+
{
|
| 1710 |
+
"epoch": 1.4896998234255445,
|
| 1711 |
+
"grad_norm": 4.39804220199585,
|
| 1712 |
+
"learning_rate": 1.4831583923105e-06,
|
| 1713 |
+
"logits/chosen": 12.365824699401855,
|
| 1714 |
+
"logits/rejected": 11.472888946533203,
|
| 1715 |
+
"logps/chosen": -4.707729816436768,
|
| 1716 |
+
"logps/rejected": -8.413005828857422,
|
| 1717 |
+
"loss": 0.2398,
|
| 1718 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 1719 |
+
"rewards/chosen": -7.061594486236572,
|
| 1720 |
+
"rewards/margins": 5.557913780212402,
|
| 1721 |
+
"rewards/rejected": -12.619508743286133,
|
| 1722 |
+
"step": 950
|
| 1723 |
+
},
|
| 1724 |
+
{
|
| 1725 |
+
"epoch": 1.4896998234255445,
|
| 1726 |
+
"eval_logits/chosen": 12.97492790222168,
|
| 1727 |
+
"eval_logits/rejected": 11.935004234313965,
|
| 1728 |
+
"eval_logps/chosen": -4.231271743774414,
|
| 1729 |
+
"eval_logps/rejected": -8.475292205810547,
|
| 1730 |
+
"eval_loss": 0.24798962473869324,
|
| 1731 |
+
"eval_rewards/accuracies": 1.0,
|
| 1732 |
+
"eval_rewards/chosen": -6.346907615661621,
|
| 1733 |
+
"eval_rewards/margins": 6.366030693054199,
|
| 1734 |
+
"eval_rewards/rejected": -12.71293830871582,
|
| 1735 |
+
"eval_runtime": 14.5034,
|
| 1736 |
+
"eval_samples_per_second": 28.407,
|
| 1737 |
+
"eval_steps_per_second": 3.585,
|
| 1738 |
+
"step": 950
|
| 1739 |
+
},
|
| 1740 |
+
{
|
| 1741 |
+
"epoch": 1.5053953305866195,
|
| 1742 |
+
"grad_norm": 2.8650925159454346,
|
| 1743 |
+
"learning_rate": 1.4355517710873184e-06,
|
| 1744 |
+
"logits/chosen": 12.443435668945312,
|
| 1745 |
+
"logits/rejected": 11.93461799621582,
|
| 1746 |
+
"logps/chosen": -4.258164405822754,
|
| 1747 |
+
"logps/rejected": -7.894736289978027,
|
| 1748 |
+
"loss": 0.1948,
|
| 1749 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 1750 |
+
"rewards/chosen": -6.387246131896973,
|
| 1751 |
+
"rewards/margins": 5.454858779907227,
|
| 1752 |
+
"rewards/rejected": -11.8421049118042,
|
| 1753 |
+
"step": 960
|
| 1754 |
+
},
|
| 1755 |
+
{
|
| 1756 |
+
"epoch": 1.5210908377476948,
|
| 1757 |
+
"grad_norm": 3.400313377380371,
|
| 1758 |
+
"learning_rate": 1.388412052037682e-06,
|
| 1759 |
+
"logits/chosen": 12.361493110656738,
|
| 1760 |
+
"logits/rejected": 11.476273536682129,
|
| 1761 |
+
"logps/chosen": -4.457276344299316,
|
| 1762 |
+
"logps/rejected": -8.228893280029297,
|
| 1763 |
+
"loss": 0.1888,
|
| 1764 |
+
"rewards/accuracies": 0.987500011920929,
|
| 1765 |
+
"rewards/chosen": -6.685914039611816,
|
| 1766 |
+
"rewards/margins": 5.657425880432129,
|
| 1767 |
+
"rewards/rejected": -12.343339920043945,
|
| 1768 |
+
"step": 970
|
| 1769 |
+
},
|
| 1770 |
+
{
|
| 1771 |
+
"epoch": 1.5367863449087698,
|
| 1772 |
+
"grad_norm": 3.8736472129821777,
|
| 1773 |
+
"learning_rate": 1.3417599122003464e-06,
|
| 1774 |
+
"logits/chosen": 12.806173324584961,
|
| 1775 |
+
"logits/rejected": 11.818658828735352,
|
| 1776 |
+
"logps/chosen": -4.827099800109863,
|
| 1777 |
+
"logps/rejected": -8.408552169799805,
|
| 1778 |
+
"loss": 0.231,
|
| 1779 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1780 |
+
"rewards/chosen": -7.240649223327637,
|
| 1781 |
+
"rewards/margins": 5.3721795082092285,
|
| 1782 |
+
"rewards/rejected": -12.612829208374023,
|
| 1783 |
+
"step": 980
|
| 1784 |
+
},
|
| 1785 |
+
{
|
| 1786 |
+
"epoch": 1.5524818520698451,
|
| 1787 |
+
"grad_norm": 3.6417574882507324,
|
| 1788 |
+
"learning_rate": 1.2956158147457116e-06,
|
| 1789 |
+
"logits/chosen": 12.428319931030273,
|
| 1790 |
+
"logits/rejected": 11.719505310058594,
|
| 1791 |
+
"logps/chosen": -5.147378444671631,
|
| 1792 |
+
"logps/rejected": -8.708709716796875,
|
| 1793 |
+
"loss": 0.2265,
|
| 1794 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1795 |
+
"rewards/chosen": -7.721067905426025,
|
| 1796 |
+
"rewards/margins": 5.341997146606445,
|
| 1797 |
+
"rewards/rejected": -13.06306266784668,
|
| 1798 |
+
"step": 990
|
| 1799 |
+
},
|
| 1800 |
+
{
|
| 1801 |
+
"epoch": 1.56817735923092,
|
| 1802 |
+
"grad_norm": 2.3367979526519775,
|
| 1803 |
+
"learning_rate": 1.2500000000000007e-06,
|
| 1804 |
+
"logits/chosen": 12.109533309936523,
|
| 1805 |
+
"logits/rejected": 11.499246597290039,
|
| 1806 |
+
"logps/chosen": -5.072439193725586,
|
| 1807 |
+
"logps/rejected": -9.098675727844238,
|
| 1808 |
+
"loss": 0.216,
|
| 1809 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 1810 |
+
"rewards/chosen": -7.608658790588379,
|
| 1811 |
+
"rewards/margins": 6.039353847503662,
|
| 1812 |
+
"rewards/rejected": -13.648012161254883,
|
| 1813 |
+
"step": 1000
|
| 1814 |
+
},
|
| 1815 |
+
{
|
| 1816 |
+
"epoch": 1.56817735923092,
|
| 1817 |
+
"eval_logits/chosen": 12.952030181884766,
|
| 1818 |
+
"eval_logits/rejected": 11.900659561157227,
|
| 1819 |
+
"eval_logps/chosen": -4.205340385437012,
|
| 1820 |
+
"eval_logps/rejected": -8.569601058959961,
|
| 1821 |
+
"eval_loss": 0.2481767237186432,
|
| 1822 |
+
"eval_rewards/accuracies": 1.0,
|
| 1823 |
+
"eval_rewards/chosen": -6.308011054992676,
|
| 1824 |
+
"eval_rewards/margins": 6.546389579772949,
|
| 1825 |
+
"eval_rewards/rejected": -12.854399681091309,
|
| 1826 |
+
"eval_runtime": 14.5049,
|
| 1827 |
+
"eval_samples_per_second": 28.404,
|
| 1828 |
+
"eval_steps_per_second": 3.585,
|
| 1829 |
+
"step": 1000
|
| 1830 |
+
},
|
| 1831 |
+
{
|
| 1832 |
+
"epoch": 1.5838728663919954,
|
| 1833 |
+
"grad_norm": 2.5456700325012207,
|
| 1834 |
+
"learning_rate": 1.204932476567175e-06,
|
| 1835 |
+
"logits/chosen": 11.960077285766602,
|
| 1836 |
+
"logits/rejected": 10.87277603149414,
|
| 1837 |
+
"logps/chosen": -4.433133125305176,
|
| 1838 |
+
"logps/rejected": -8.525662422180176,
|
| 1839 |
+
"loss": 0.2443,
|
| 1840 |
+
"rewards/accuracies": 0.987500011920929,
|
| 1841 |
+
"rewards/chosen": -6.6496992111206055,
|
| 1842 |
+
"rewards/margins": 6.1387939453125,
|
| 1843 |
+
"rewards/rejected": -12.788493156433105,
|
| 1844 |
+
"step": 1010
|
| 1845 |
+
},
|
| 1846 |
+
{
|
| 1847 |
+
"epoch": 1.5995683735530704,
|
| 1848 |
+
"grad_norm": 2.834538698196411,
|
| 1849 |
+
"learning_rate": 1.160433012552508e-06,
|
| 1850 |
+
"logits/chosen": 12.570700645446777,
|
| 1851 |
+
"logits/rejected": 11.804948806762695,
|
| 1852 |
+
"logps/chosen": -3.698901414871216,
|
| 1853 |
+
"logps/rejected": -7.383004665374756,
|
| 1854 |
+
"loss": 0.2226,
|
| 1855 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1856 |
+
"rewards/chosen": -5.5483527183532715,
|
| 1857 |
+
"rewards/margins": 5.526154518127441,
|
| 1858 |
+
"rewards/rejected": -11.074506759643555,
|
| 1859 |
+
"step": 1020
|
| 1860 |
+
},
|
| 1861 |
+
{
|
| 1862 |
+
"epoch": 1.6152638807141457,
|
| 1863 |
+
"grad_norm": 3.336750030517578,
|
| 1864 |
+
"learning_rate": 1.11652112689164e-06,
|
| 1865 |
+
"logits/chosen": 12.501141548156738,
|
| 1866 |
+
"logits/rejected": 11.617207527160645,
|
| 1867 |
+
"logps/chosen": -4.3009490966796875,
|
| 1868 |
+
"logps/rejected": -8.17215347290039,
|
| 1869 |
+
"loss": 0.2303,
|
| 1870 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 1871 |
+
"rewards/chosen": -6.451423645019531,
|
| 1872 |
+
"rewards/margins": 5.806807041168213,
|
| 1873 |
+
"rewards/rejected": -12.258230209350586,
|
| 1874 |
+
"step": 1030
|
| 1875 |
+
},
|
| 1876 |
+
{
|
| 1877 |
+
"epoch": 1.6309593878752207,
|
| 1878 |
+
"grad_norm": 2.519408941268921,
|
| 1879 |
+
"learning_rate": 1.073216080788921e-06,
|
| 1880 |
+
"logits/chosen": 12.614023208618164,
|
| 1881 |
+
"logits/rejected": 11.761343002319336,
|
| 1882 |
+
"logps/chosen": -4.564716815948486,
|
| 1883 |
+
"logps/rejected": -8.430013656616211,
|
| 1884 |
+
"loss": 0.2634,
|
| 1885 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 1886 |
+
"rewards/chosen": -6.847075462341309,
|
| 1887 |
+
"rewards/margins": 5.797944068908691,
|
| 1888 |
+
"rewards/rejected": -12.645018577575684,
|
| 1889 |
+
"step": 1040
|
| 1890 |
+
},
|
| 1891 |
+
{
|
| 1892 |
+
"epoch": 1.6466548950362958,
|
| 1893 |
+
"grad_norm": 4.510939121246338,
|
| 1894 |
+
"learning_rate": 1.0305368692688175e-06,
|
| 1895 |
+
"logits/chosen": 12.286882400512695,
|
| 1896 |
+
"logits/rejected": 11.735586166381836,
|
| 1897 |
+
"logps/chosen": -4.190676212310791,
|
| 1898 |
+
"logps/rejected": -8.055787086486816,
|
| 1899 |
+
"loss": 0.1713,
|
| 1900 |
+
"rewards/accuracies": 0.9375,
|
| 1901 |
+
"rewards/chosen": -6.286014556884766,
|
| 1902 |
+
"rewards/margins": 5.797664642333984,
|
| 1903 |
+
"rewards/rejected": -12.08367919921875,
|
| 1904 |
+
"step": 1050
|
| 1905 |
+
},
|
| 1906 |
+
{
|
| 1907 |
+
"epoch": 1.6466548950362958,
|
| 1908 |
+
"eval_logits/chosen": 12.861144065856934,
|
| 1909 |
+
"eval_logits/rejected": 11.838491439819336,
|
| 1910 |
+
"eval_logps/chosen": -4.190127372741699,
|
| 1911 |
+
"eval_logps/rejected": -8.606769561767578,
|
| 1912 |
+
"eval_loss": 0.24250496923923492,
|
| 1913 |
+
"eval_rewards/accuracies": 1.0,
|
| 1914 |
+
"eval_rewards/chosen": -6.285191535949707,
|
| 1915 |
+
"eval_rewards/margins": 6.624964237213135,
|
| 1916 |
+
"eval_rewards/rejected": -12.910155296325684,
|
| 1917 |
+
"eval_runtime": 14.4989,
|
| 1918 |
+
"eval_samples_per_second": 28.416,
|
| 1919 |
+
"eval_steps_per_second": 3.586,
|
| 1920 |
+
"step": 1050
|
| 1921 |
+
},
|
| 1922 |
+
{
|
| 1923 |
+
"epoch": 1.662350402197371,
|
| 1924 |
+
"grad_norm": 3.669426441192627,
|
| 1925 |
+
"learning_rate": 9.88502212844063e-07,
|
| 1926 |
+
"logits/chosen": 12.255383491516113,
|
| 1927 |
+
"logits/rejected": 11.495210647583008,
|
| 1928 |
+
"logps/chosen": -3.948782444000244,
|
| 1929 |
+
"logps/rejected": -8.229391098022461,
|
| 1930 |
+
"loss": 0.2143,
|
| 1931 |
+
"rewards/accuracies": 0.987500011920929,
|
| 1932 |
+
"rewards/chosen": -5.923173427581787,
|
| 1933 |
+
"rewards/margins": 6.4209136962890625,
|
| 1934 |
+
"rewards/rejected": -12.344087600708008,
|
| 1935 |
+
"step": 1060
|
| 1936 |
+
},
|
| 1937 |
+
{
|
| 1938 |
+
"epoch": 1.678045909358446,
|
| 1939 |
+
"grad_norm": 4.005455493927002,
|
| 1940 |
+
"learning_rate": 9.471305493042243e-07,
|
| 1941 |
+
"logits/chosen": 12.1824312210083,
|
| 1942 |
+
"logits/rejected": 11.290962219238281,
|
| 1943 |
+
"logps/chosen": -4.578344821929932,
|
| 1944 |
+
"logps/rejected": -8.505768775939941,
|
| 1945 |
+
"loss": 0.2361,
|
| 1946 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1947 |
+
"rewards/chosen": -6.867516994476318,
|
| 1948 |
+
"rewards/margins": 5.89113712310791,
|
| 1949 |
+
"rewards/rejected": -12.75865364074707,
|
| 1950 |
+
"step": 1070
|
| 1951 |
+
},
|
| 1952 |
+
{
|
| 1953 |
+
"epoch": 1.6937414165195213,
|
| 1954 |
+
"grad_norm": 2.032545566558838,
|
| 1955 |
+
"learning_rate": 9.064400256282757e-07,
|
| 1956 |
+
"logits/chosen": 12.451539039611816,
|
| 1957 |
+
"logits/rejected": 11.746088027954102,
|
| 1958 |
+
"logps/chosen": -4.1952643394470215,
|
| 1959 |
+
"logps/rejected": -7.86816930770874,
|
| 1960 |
+
"loss": 0.1957,
|
| 1961 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1962 |
+
"rewards/chosen": -6.2928972244262695,
|
| 1963 |
+
"rewards/margins": 5.50935697555542,
|
| 1964 |
+
"rewards/rejected": -11.802255630493164,
|
| 1965 |
+
"step": 1080
|
| 1966 |
+
},
|
| 1967 |
+
{
|
| 1968 |
+
"epoch": 1.7094369236805964,
|
| 1969 |
+
"grad_norm": 4.466379642486572,
|
| 1970 |
+
"learning_rate": 8.664484900247363e-07,
|
| 1971 |
+
"logits/chosen": 12.144006729125977,
|
| 1972 |
+
"logits/rejected": 11.437246322631836,
|
| 1973 |
+
"logps/chosen": -4.593950271606445,
|
| 1974 |
+
"logps/rejected": -8.530038833618164,
|
| 1975 |
+
"loss": 0.186,
|
| 1976 |
+
"rewards/accuracies": 0.9375,
|
| 1977 |
+
"rewards/chosen": -6.890925407409668,
|
| 1978 |
+
"rewards/margins": 5.9041337966918945,
|
| 1979 |
+
"rewards/rejected": -12.795059204101562,
|
| 1980 |
+
"step": 1090
|
| 1981 |
+
},
|
| 1982 |
+
{
|
| 1983 |
+
"epoch": 1.7251324308416716,
|
| 1984 |
+
"grad_norm": 2.941102981567383,
|
| 1985 |
+
"learning_rate": 8.271734841028553e-07,
|
| 1986 |
+
"logits/chosen": 12.305268287658691,
|
| 1987 |
+
"logits/rejected": 11.327041625976562,
|
| 1988 |
+
"logps/chosen": -4.72153377532959,
|
| 1989 |
+
"logps/rejected": -8.988466262817383,
|
| 1990 |
+
"loss": 0.2264,
|
| 1991 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 1992 |
+
"rewards/chosen": -7.082299709320068,
|
| 1993 |
+
"rewards/margins": 6.400399684906006,
|
| 1994 |
+
"rewards/rejected": -13.482701301574707,
|
| 1995 |
+
"step": 1100
|
| 1996 |
+
},
|
| 1997 |
+
{
|
| 1998 |
+
"epoch": 1.7251324308416716,
|
| 1999 |
+
"eval_logits/chosen": 12.896148681640625,
|
| 2000 |
+
"eval_logits/rejected": 11.868448257446289,
|
| 2001 |
+
"eval_logps/chosen": -4.222008228302002,
|
| 2002 |
+
"eval_logps/rejected": -8.642206192016602,
|
| 2003 |
+
"eval_loss": 0.2400711625814438,
|
| 2004 |
+
"eval_rewards/accuracies": 1.0,
|
| 2005 |
+
"eval_rewards/chosen": -6.333012580871582,
|
| 2006 |
+
"eval_rewards/margins": 6.6302971839904785,
|
| 2007 |
+
"eval_rewards/rejected": -12.963308334350586,
|
| 2008 |
+
"eval_runtime": 14.5019,
|
| 2009 |
+
"eval_samples_per_second": 28.41,
|
| 2010 |
+
"eval_steps_per_second": 3.586,
|
| 2011 |
+
"step": 1100
|
| 2012 |
+
},
|
| 2013 |
+
{
|
| 2014 |
+
"epoch": 1.7408279380027467,
|
| 2015 |
+
"grad_norm": 4.001830101013184,
|
| 2016 |
+
"learning_rate": 7.886322351782782e-07,
|
| 2017 |
+
"logits/chosen": 12.858680725097656,
|
| 2018 |
+
"logits/rejected": 11.653299331665039,
|
| 2019 |
+
"logps/chosen": -4.239178657531738,
|
| 2020 |
+
"logps/rejected": -8.762260437011719,
|
| 2021 |
+
"loss": 0.2786,
|
| 2022 |
+
"rewards/accuracies": 0.9375,
|
| 2023 |
+
"rewards/chosen": -6.358767986297607,
|
| 2024 |
+
"rewards/margins": 6.784623622894287,
|
| 2025 |
+
"rewards/rejected": -13.143391609191895,
|
| 2026 |
+
"step": 1110
|
| 2027 |
+
},
|
| 2028 |
+
{
|
| 2029 |
+
"epoch": 1.7565234451638219,
|
| 2030 |
+
"grad_norm": 2.9689064025878906,
|
| 2031 |
+
"learning_rate": 7.508416487165862e-07,
|
| 2032 |
+
"logits/chosen": 12.076465606689453,
|
| 2033 |
+
"logits/rejected": 11.392438888549805,
|
| 2034 |
+
"logps/chosen": -4.560977458953857,
|
| 2035 |
+
"logps/rejected": -8.554190635681152,
|
| 2036 |
+
"loss": 0.1998,
|
| 2037 |
+
"rewards/accuracies": 0.9375,
|
| 2038 |
+
"rewards/chosen": -6.841466426849365,
|
| 2039 |
+
"rewards/margins": 5.98982048034668,
|
| 2040 |
+
"rewards/rejected": -12.831286430358887,
|
| 2041 |
+
"step": 1120
|
| 2042 |
+
},
|
| 2043 |
+
{
|
| 2044 |
+
"epoch": 1.772218952324897,
|
| 2045 |
+
"grad_norm": 1.6012306213378906,
|
| 2046 |
+
"learning_rate": 7.138183009179922e-07,
|
| 2047 |
+
"logits/chosen": 12.253329277038574,
|
| 2048 |
+
"logits/rejected": 11.584481239318848,
|
| 2049 |
+
"logps/chosen": -4.0213422775268555,
|
| 2050 |
+
"logps/rejected": -8.13912296295166,
|
| 2051 |
+
"loss": 0.1823,
|
| 2052 |
+
"rewards/accuracies": 0.949999988079071,
|
| 2053 |
+
"rewards/chosen": -6.032013893127441,
|
| 2054 |
+
"rewards/margins": 6.176670074462891,
|
| 2055 |
+
"rewards/rejected": -12.208683967590332,
|
| 2056 |
+
"step": 1130
|
| 2057 |
+
},
|
| 2058 |
+
{
|
| 2059 |
+
"epoch": 1.7879144594859722,
|
| 2060 |
+
"grad_norm": 3.8666510581970215,
|
| 2061 |
+
"learning_rate": 6.775784314464717e-07,
|
| 2062 |
+
"logits/chosen": 11.831202507019043,
|
| 2063 |
+
"logits/rejected": 11.49097728729248,
|
| 2064 |
+
"logps/chosen": -4.555119037628174,
|
| 2065 |
+
"logps/rejected": -8.43702220916748,
|
| 2066 |
+
"loss": 0.2494,
|
| 2067 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 2068 |
+
"rewards/chosen": -6.832678318023682,
|
| 2069 |
+
"rewards/margins": 5.822856426239014,
|
| 2070 |
+
"rewards/rejected": -12.655534744262695,
|
| 2071 |
+
"step": 1140
|
| 2072 |
+
},
|
| 2073 |
+
{
|
| 2074 |
+
"epoch": 1.8036099666470473,
|
| 2075 |
+
"grad_norm": 3.794703483581543,
|
| 2076 |
+
"learning_rate": 6.421379363065142e-07,
|
| 2077 |
+
"logits/chosen": 12.098966598510742,
|
| 2078 |
+
"logits/rejected": 11.32978343963623,
|
| 2079 |
+
"logps/chosen": -4.038215637207031,
|
| 2080 |
+
"logps/rejected": -7.658058166503906,
|
| 2081 |
+
"loss": 0.2365,
|
| 2082 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 2083 |
+
"rewards/chosen": -6.057322978973389,
|
| 2084 |
+
"rewards/margins": 5.4297637939453125,
|
| 2085 |
+
"rewards/rejected": -11.487086296081543,
|
| 2086 |
+
"step": 1150
|
| 2087 |
+
},
|
| 2088 |
+
{
|
| 2089 |
+
"epoch": 1.8036099666470473,
|
| 2090 |
+
"eval_logits/chosen": 12.860236167907715,
|
| 2091 |
+
"eval_logits/rejected": 11.83856201171875,
|
| 2092 |
+
"eval_logps/chosen": -4.173120498657227,
|
| 2093 |
+
"eval_logps/rejected": -8.646206855773926,
|
| 2094 |
+
"eval_loss": 0.23765961825847626,
|
| 2095 |
+
"eval_rewards/accuracies": 1.0,
|
| 2096 |
+
"eval_rewards/chosen": -6.2596821784973145,
|
| 2097 |
+
"eval_rewards/margins": 6.709629535675049,
|
| 2098 |
+
"eval_rewards/rejected": -12.96930980682373,
|
| 2099 |
+
"eval_runtime": 14.5141,
|
| 2100 |
+
"eval_samples_per_second": 28.386,
|
| 2101 |
+
"eval_steps_per_second": 3.583,
|
| 2102 |
+
"step": 1150
|
| 2103 |
+
},
|
| 2104 |
+
{
|
| 2105 |
+
"epoch": 1.8193054738081225,
|
| 2106 |
+
"grad_norm": 4.030630588531494,
|
| 2107 |
+
"learning_rate": 6.075123608706093e-07,
|
| 2108 |
+
"logits/chosen": 12.267806053161621,
|
| 2109 |
+
"logits/rejected": 11.600652694702148,
|
| 2110 |
+
"logps/chosen": -4.568865776062012,
|
| 2111 |
+
"logps/rejected": -7.858088493347168,
|
| 2112 |
+
"loss": 0.2043,
|
| 2113 |
+
"rewards/accuracies": 0.949999988079071,
|
| 2114 |
+
"rewards/chosen": -6.853298187255859,
|
| 2115 |
+
"rewards/margins": 4.933834552764893,
|
| 2116 |
+
"rewards/rejected": -11.78713321685791,
|
| 2117 |
+
"step": 1160
|
| 2118 |
+
},
|
| 2119 |
+
{
|
| 2120 |
+
"epoch": 1.8350009809691976,
|
| 2121 |
+
"grad_norm": 3.0004727840423584,
|
| 2122 |
+
"learning_rate": 5.737168930605272e-07,
|
| 2123 |
+
"logits/chosen": 12.178464889526367,
|
| 2124 |
+
"logits/rejected": 11.31472110748291,
|
| 2125 |
+
"logps/chosen": -5.0792646408081055,
|
| 2126 |
+
"logps/rejected": -8.85279369354248,
|
| 2127 |
+
"loss": 0.1923,
|
| 2128 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 2129 |
+
"rewards/chosen": -7.618897438049316,
|
| 2130 |
+
"rewards/margins": 5.6602935791015625,
|
| 2131 |
+
"rewards/rejected": -13.279190063476562,
|
| 2132 |
+
"step": 1170
|
| 2133 |
+
},
|
| 2134 |
+
{
|
| 2135 |
+
"epoch": 1.8506964881302728,
|
| 2136 |
+
"grad_norm": 3.7206344604492188,
|
| 2137 |
+
"learning_rate": 5.407663566854008e-07,
|
| 2138 |
+
"logits/chosen": 12.098569869995117,
|
| 2139 |
+
"logits/rejected": 11.365894317626953,
|
| 2140 |
+
"logps/chosen": -5.336371898651123,
|
| 2141 |
+
"logps/rejected": -9.606247901916504,
|
| 2142 |
+
"loss": 0.2206,
|
| 2143 |
+
"rewards/accuracies": 0.987500011920929,
|
| 2144 |
+
"rewards/chosen": -8.004558563232422,
|
| 2145 |
+
"rewards/margins": 6.404813289642334,
|
| 2146 |
+
"rewards/rejected": -14.409372329711914,
|
| 2147 |
+
"step": 1180
|
| 2148 |
+
},
|
| 2149 |
+
{
|
| 2150 |
+
"epoch": 1.8663919952913477,
|
| 2151 |
+
"grad_norm": 4.599307537078857,
|
| 2152 |
+
"learning_rate": 5.086752049395094e-07,
|
| 2153 |
+
"logits/chosen": 12.062299728393555,
|
| 2154 |
+
"logits/rejected": 11.360358238220215,
|
| 2155 |
+
"logps/chosen": -5.103975772857666,
|
| 2156 |
+
"logps/rejected": -9.448270797729492,
|
| 2157 |
+
"loss": 0.2338,
|
| 2158 |
+
"rewards/accuracies": 0.925000011920929,
|
| 2159 |
+
"rewards/chosen": -7.655962944030762,
|
| 2160 |
+
"rewards/margins": 6.516444206237793,
|
| 2161 |
+
"rewards/rejected": -14.172407150268555,
|
| 2162 |
+
"step": 1190
|
| 2163 |
+
},
|
| 2164 |
+
{
|
| 2165 |
+
"epoch": 1.882087502452423,
|
| 2166 |
+
"grad_norm": 2.590059757232666,
|
| 2167 |
+
"learning_rate": 4.774575140626317e-07,
|
| 2168 |
+
"logits/chosen": 12.54432201385498,
|
| 2169 |
+
"logits/rejected": 11.949193000793457,
|
| 2170 |
+
"logps/chosen": -4.637765884399414,
|
| 2171 |
+
"logps/rejected": -8.537626266479492,
|
| 2172 |
+
"loss": 0.2028,
|
| 2173 |
+
"rewards/accuracies": 0.9375,
|
| 2174 |
+
"rewards/chosen": -6.956648826599121,
|
| 2175 |
+
"rewards/margins": 5.849791526794434,
|
| 2176 |
+
"rewards/rejected": -12.806440353393555,
|
| 2177 |
+
"step": 1200
|
| 2178 |
+
},
|
| 2179 |
+
{
|
| 2180 |
+
"epoch": 1.882087502452423,
|
| 2181 |
+
"eval_logits/chosen": 12.827628135681152,
|
| 2182 |
+
"eval_logits/rejected": 11.809093475341797,
|
| 2183 |
+
"eval_logps/chosen": -4.21560001373291,
|
| 2184 |
+
"eval_logps/rejected": -8.739953994750977,
|
| 2185 |
+
"eval_loss": 0.23520438373088837,
|
| 2186 |
+
"eval_rewards/accuracies": 1.0,
|
| 2187 |
+
"eval_rewards/chosen": -6.323400497436523,
|
| 2188 |
+
"eval_rewards/margins": 6.786531448364258,
|
| 2189 |
+
"eval_rewards/rejected": -13.109930992126465,
|
| 2190 |
+
"eval_runtime": 14.5066,
|
| 2191 |
+
"eval_samples_per_second": 28.401,
|
| 2192 |
+
"eval_steps_per_second": 3.585,
|
| 2193 |
+
"step": 1200
|
| 2194 |
+
},
|
| 2195 |
+
{
|
| 2196 |
+
"epoch": 1.897783009613498,
|
| 2197 |
+
"grad_norm": 2.3263192176818848,
|
| 2198 |
+
"learning_rate": 4.4712697716573994e-07,
|
| 2199 |
+
"logits/chosen": 11.775550842285156,
|
| 2200 |
+
"logits/rejected": 11.042590141296387,
|
| 2201 |
+
"logps/chosen": -5.221895694732666,
|
| 2202 |
+
"logps/rejected": -9.288631439208984,
|
| 2203 |
+
"loss": 0.2082,
|
| 2204 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 2205 |
+
"rewards/chosen": -7.832842826843262,
|
| 2206 |
+
"rewards/margins": 6.100103855133057,
|
| 2207 |
+
"rewards/rejected": -13.932948112487793,
|
| 2208 |
+
"step": 1210
|
| 2209 |
+
},
|
| 2210 |
+
{
|
| 2211 |
+
"epoch": 1.9134785167745734,
|
| 2212 |
+
"grad_norm": 3.9693384170532227,
|
| 2213 |
+
"learning_rate": 4.1769689822475147e-07,
|
| 2214 |
+
"logits/chosen": 12.233749389648438,
|
| 2215 |
+
"logits/rejected": 11.018649101257324,
|
| 2216 |
+
"logps/chosen": -4.284747123718262,
|
| 2217 |
+
"logps/rejected": -8.676647186279297,
|
| 2218 |
+
"loss": 0.2287,
|
| 2219 |
+
"rewards/accuracies": 0.949999988079071,
|
| 2220 |
+
"rewards/chosen": -6.427121162414551,
|
| 2221 |
+
"rewards/margins": 6.587851047515869,
|
| 2222 |
+
"rewards/rejected": -13.014971733093262,
|
| 2223 |
+
"step": 1220
|
| 2224 |
+
},
|
| 2225 |
+
{
|
| 2226 |
+
"epoch": 1.9291740239356483,
|
| 2227 |
+
"grad_norm": 3.8713510036468506,
|
| 2228 |
+
"learning_rate": 3.891801862449629e-07,
|
| 2229 |
+
"logits/chosen": 12.066610336303711,
|
| 2230 |
+
"logits/rejected": 11.343181610107422,
|
| 2231 |
+
"logps/chosen": -4.260552883148193,
|
| 2232 |
+
"logps/rejected": -8.419907569885254,
|
| 2233 |
+
"loss": 0.2681,
|
| 2234 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 2235 |
+
"rewards/chosen": -6.390830039978027,
|
| 2236 |
+
"rewards/margins": 6.239031791687012,
|
| 2237 |
+
"rewards/rejected": -12.629861831665039,
|
| 2238 |
+
"step": 1230
|
| 2239 |
+
},
|
| 2240 |
+
{
|
| 2241 |
+
"epoch": 1.9448695310967237,
|
| 2242 |
+
"grad_norm": 3.3727104663848877,
|
| 2243 |
+
"learning_rate": 3.615893495987335e-07,
|
| 2244 |
+
"logits/chosen": 12.09786605834961,
|
| 2245 |
+
"logits/rejected": 10.789514541625977,
|
| 2246 |
+
"logps/chosen": -4.765267372131348,
|
| 2247 |
+
"logps/rejected": -8.926709175109863,
|
| 2248 |
+
"loss": 0.213,
|
| 2249 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 2250 |
+
"rewards/chosen": -7.1479010581970215,
|
| 2251 |
+
"rewards/margins": 6.242161750793457,
|
| 2252 |
+
"rewards/rejected": -13.390063285827637,
|
| 2253 |
+
"step": 1240
|
| 2254 |
+
},
|
| 2255 |
+
{
|
| 2256 |
+
"epoch": 1.9605650382577986,
|
| 2257 |
+
"grad_norm": 5.0855255126953125,
|
| 2258 |
+
"learning_rate": 3.3493649053890325e-07,
|
| 2259 |
+
"logits/chosen": 12.1903715133667,
|
| 2260 |
+
"logits/rejected": 11.303841590881348,
|
| 2261 |
+
"logps/chosen": -4.052525997161865,
|
| 2262 |
+
"logps/rejected": -8.359174728393555,
|
| 2263 |
+
"loss": 0.2177,
|
| 2264 |
+
"rewards/accuracies": 0.987500011920929,
|
| 2265 |
+
"rewards/chosen": -6.078789710998535,
|
| 2266 |
+
"rewards/margins": 6.4599738121032715,
|
| 2267 |
+
"rewards/rejected": -12.538763046264648,
|
| 2268 |
+
"step": 1250
|
| 2269 |
+
},
|
| 2270 |
+
{
|
| 2271 |
+
"epoch": 1.9605650382577986,
|
| 2272 |
+
"eval_logits/chosen": 12.821044921875,
|
| 2273 |
+
"eval_logits/rejected": 11.804606437683105,
|
| 2274 |
+
"eval_logps/chosen": -4.163019180297852,
|
| 2275 |
+
"eval_logps/rejected": -8.714776039123535,
|
| 2276 |
+
"eval_loss": 0.23355062305927277,
|
| 2277 |
+
"eval_rewards/accuracies": 1.0,
|
| 2278 |
+
"eval_rewards/chosen": -6.244529724121094,
|
| 2279 |
+
"eval_rewards/margins": 6.827635765075684,
|
| 2280 |
+
"eval_rewards/rejected": -13.072164535522461,
|
| 2281 |
+
"eval_runtime": 14.5059,
|
| 2282 |
+
"eval_samples_per_second": 28.402,
|
| 2283 |
+
"eval_steps_per_second": 3.585,
|
| 2284 |
+
"step": 1250
|
| 2285 |
+
},
|
| 2286 |
+
{
|
| 2287 |
+
"epoch": 1.976260545418874,
|
| 2288 |
+
"grad_norm": 2.8941688537597656,
|
| 2289 |
+
"learning_rate": 3.092332998903416e-07,
|
| 2290 |
+
"logits/chosen": 12.222402572631836,
|
| 2291 |
+
"logits/rejected": 11.670201301574707,
|
| 2292 |
+
"logps/chosen": -4.20303201675415,
|
| 2293 |
+
"logps/rejected": -7.536846160888672,
|
| 2294 |
+
"loss": 0.2015,
|
| 2295 |
+
"rewards/accuracies": 0.925000011920929,
|
| 2296 |
+
"rewards/chosen": -6.3045477867126465,
|
| 2297 |
+
"rewards/margins": 5.000721454620361,
|
| 2298 |
+
"rewards/rejected": -11.305269241333008,
|
| 2299 |
+
"step": 1260
|
| 2300 |
+
},
|
| 2301 |
+
{
|
| 2302 |
+
"epoch": 1.991956052579949,
|
| 2303 |
+
"grad_norm": 5.021040439605713,
|
| 2304 |
+
"learning_rate": 2.844910519219632e-07,
|
| 2305 |
+
"logits/chosen": 11.843130111694336,
|
| 2306 |
+
"logits/rejected": 11.057130813598633,
|
| 2307 |
+
"logps/chosen": -4.563799858093262,
|
| 2308 |
+
"logps/rejected": -8.902048110961914,
|
| 2309 |
+
"loss": 0.238,
|
| 2310 |
+
"rewards/accuracies": 0.9375,
|
| 2311 |
+
"rewards/chosen": -6.845700740814209,
|
| 2312 |
+
"rewards/margins": 6.507371425628662,
|
| 2313 |
+
"rewards/rejected": -13.353071212768555,
|
| 2314 |
+
"step": 1270
|
| 2315 |
+
},
|
| 2316 |
+
{
|
| 2317 |
+
"epoch": 2.00627820286443,
|
| 2318 |
+
"grad_norm": 2.88142466545105,
|
| 2319 |
+
"learning_rate": 2.6072059940146775e-07,
|
| 2320 |
+
"logits/chosen": 12.205242156982422,
|
| 2321 |
+
"logits/rejected": 11.59632396697998,
|
| 2322 |
+
"logps/chosen": -4.475119590759277,
|
| 2323 |
+
"logps/rejected": -8.82370662689209,
|
| 2324 |
+
"loss": 0.1594,
|
| 2325 |
+
"rewards/accuracies": 0.9726027250289917,
|
| 2326 |
+
"rewards/chosen": -6.712679386138916,
|
| 2327 |
+
"rewards/margins": 6.522879600524902,
|
| 2328 |
+
"rewards/rejected": -13.235559463500977,
|
| 2329 |
+
"step": 1280
|
| 2330 |
+
},
|
| 2331 |
+
{
|
| 2332 |
+
"epoch": 2.021973710025505,
|
| 2333 |
+
"grad_norm": 4.564095973968506,
|
| 2334 |
+
"learning_rate": 2.3793236883495164e-07,
|
| 2335 |
+
"logits/chosen": 12.256528854370117,
|
| 2336 |
+
"logits/rejected": 11.842262268066406,
|
| 2337 |
+
"logps/chosen": -4.68477725982666,
|
| 2338 |
+
"logps/rejected": -8.527600288391113,
|
| 2339 |
+
"loss": 0.2518,
|
| 2340 |
+
"rewards/accuracies": 0.925000011920929,
|
| 2341 |
+
"rewards/chosen": -7.027165412902832,
|
| 2342 |
+
"rewards/margins": 5.764235496520996,
|
| 2343 |
+
"rewards/rejected": -12.791400909423828,
|
| 2344 |
+
"step": 1290
|
| 2345 |
+
},
|
| 2346 |
+
{
|
| 2347 |
+
"epoch": 2.0376692171865805,
|
| 2348 |
+
"grad_norm": 2.108403444290161,
|
| 2349 |
+
"learning_rate": 2.1613635589349756e-07,
|
| 2350 |
+
"logits/chosen": 12.441131591796875,
|
| 2351 |
+
"logits/rejected": 11.819371223449707,
|
| 2352 |
+
"logps/chosen": -4.156233310699463,
|
| 2353 |
+
"logps/rejected": -7.932003974914551,
|
| 2354 |
+
"loss": 0.1959,
|
| 2355 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 2356 |
+
"rewards/chosen": -6.234350204467773,
|
| 2357 |
+
"rewards/margins": 5.663657188415527,
|
| 2358 |
+
"rewards/rejected": -11.8980073928833,
|
| 2359 |
+
"step": 1300
|
| 2360 |
+
},
|
| 2361 |
+
{
|
| 2362 |
+
"epoch": 2.0376692171865805,
|
| 2363 |
+
"eval_logits/chosen": 12.82050609588623,
|
| 2364 |
+
"eval_logits/rejected": 11.804505348205566,
|
| 2365 |
+
"eval_logps/chosen": -4.199222087860107,
|
| 2366 |
+
"eval_logps/rejected": -8.762627601623535,
|
| 2367 |
+
"eval_loss": 0.233205184340477,
|
| 2368 |
+
"eval_rewards/accuracies": 1.0,
|
| 2369 |
+
"eval_rewards/chosen": -6.29883337020874,
|
| 2370 |
+
"eval_rewards/margins": 6.845107078552246,
|
| 2371 |
+
"eval_rewards/rejected": -13.143939971923828,
|
| 2372 |
+
"eval_runtime": 14.5006,
|
| 2373 |
+
"eval_samples_per_second": 28.413,
|
| 2374 |
+
"eval_steps_per_second": 3.586,
|
| 2375 |
+
"step": 1300
|
| 2376 |
+
},
|
| 2377 |
+
{
|
| 2378 |
+
"epoch": 2.0533647243476554,
|
| 2379 |
+
"grad_norm": 2.8547580242156982,
|
| 2380 |
+
"learning_rate": 1.95342121028749e-07,
|
| 2381 |
+
"logits/chosen": 12.230999946594238,
|
| 2382 |
+
"logits/rejected": 11.622830390930176,
|
| 2383 |
+
"logps/chosen": -4.15807580947876,
|
| 2384 |
+
"logps/rejected": -8.706338882446289,
|
| 2385 |
+
"loss": 0.2139,
|
| 2386 |
+
"rewards/accuracies": 0.949999988079071,
|
| 2387 |
+
"rewards/chosen": -6.237114429473877,
|
| 2388 |
+
"rewards/margins": 6.822393894195557,
|
| 2389 |
+
"rewards/rejected": -13.059507369995117,
|
| 2390 |
+
"step": 1310
|
| 2391 |
+
},
|
| 2392 |
+
{
|
| 2393 |
+
"epoch": 2.0690602315087308,
|
| 2394 |
+
"grad_norm": 3.201754331588745,
|
| 2395 |
+
"learning_rate": 1.7555878527937164e-07,
|
| 2396 |
+
"logits/chosen": 11.826972007751465,
|
| 2397 |
+
"logits/rejected": 11.33891773223877,
|
| 2398 |
+
"logps/chosen": -4.819631099700928,
|
| 2399 |
+
"logps/rejected": -8.871285438537598,
|
| 2400 |
+
"loss": 0.1976,
|
| 2401 |
+
"rewards/accuracies": 0.949999988079071,
|
| 2402 |
+
"rewards/chosen": -7.229445457458496,
|
| 2403 |
+
"rewards/margins": 6.077480792999268,
|
| 2404 |
+
"rewards/rejected": -13.306925773620605,
|
| 2405 |
+
"step": 1320
|
| 2406 |
+
},
|
| 2407 |
+
{
|
| 2408 |
+
"epoch": 2.0847557386698057,
|
| 2409 |
+
"grad_norm": 3.492197275161743,
|
| 2410 |
+
"learning_rate": 1.567950262702714e-07,
|
| 2411 |
+
"logits/chosen": 12.089580535888672,
|
| 2412 |
+
"logits/rejected": 11.315408706665039,
|
| 2413 |
+
"logps/chosen": -4.63484525680542,
|
| 2414 |
+
"logps/rejected": -8.714259147644043,
|
| 2415 |
+
"loss": 0.2012,
|
| 2416 |
+
"rewards/accuracies": 0.949999988079071,
|
| 2417 |
+
"rewards/chosen": -6.952267646789551,
|
| 2418 |
+
"rewards/margins": 6.119122505187988,
|
| 2419 |
+
"rewards/rejected": -13.071390151977539,
|
| 2420 |
+
"step": 1330
|
| 2421 |
+
},
|
| 2422 |
+
{
|
| 2423 |
+
"epoch": 2.100451245830881,
|
| 2424 |
+
"grad_norm": 3.4347550868988037,
|
| 2425 |
+
"learning_rate": 1.3905907440629752e-07,
|
| 2426 |
+
"logits/chosen": 12.134693145751953,
|
| 2427 |
+
"logits/rejected": 10.963116645812988,
|
| 2428 |
+
"logps/chosen": -4.529653072357178,
|
| 2429 |
+
"logps/rejected": -9.13645076751709,
|
| 2430 |
+
"loss": 0.2045,
|
| 2431 |
+
"rewards/accuracies": 0.949999988079071,
|
| 2432 |
+
"rewards/chosen": -6.794480323791504,
|
| 2433 |
+
"rewards/margins": 6.9101972579956055,
|
| 2434 |
+
"rewards/rejected": -13.704676628112793,
|
| 2435 |
+
"step": 1340
|
| 2436 |
+
},
|
| 2437 |
+
{
|
| 2438 |
+
"epoch": 2.116146752991956,
|
| 2439 |
+
"grad_norm": 1.7697747945785522,
|
| 2440 |
+
"learning_rate": 1.223587092621162e-07,
|
| 2441 |
+
"logits/chosen": 12.274737358093262,
|
| 2442 |
+
"logits/rejected": 11.442537307739258,
|
| 2443 |
+
"logps/chosen": -4.994804382324219,
|
| 2444 |
+
"logps/rejected": -9.549346923828125,
|
| 2445 |
+
"loss": 0.1997,
|
| 2446 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 2447 |
+
"rewards/chosen": -7.492206573486328,
|
| 2448 |
+
"rewards/margins": 6.831814765930176,
|
| 2449 |
+
"rewards/rejected": -14.324020385742188,
|
| 2450 |
+
"step": 1350
|
| 2451 |
+
},
|
| 2452 |
+
{
|
| 2453 |
+
"epoch": 2.116146752991956,
|
| 2454 |
+
"eval_logits/chosen": 12.819293975830078,
|
| 2455 |
+
"eval_logits/rejected": 11.81404972076416,
|
| 2456 |
+
"eval_logps/chosen": -4.241227149963379,
|
| 2457 |
+
"eval_logps/rejected": -8.782449722290039,
|
| 2458 |
+
"eval_loss": 0.22972312569618225,
|
| 2459 |
+
"eval_rewards/accuracies": 1.0,
|
| 2460 |
+
"eval_rewards/chosen": -6.361841201782227,
|
| 2461 |
+
"eval_rewards/margins": 6.811832904815674,
|
| 2462 |
+
"eval_rewards/rejected": -13.173673629760742,
|
| 2463 |
+
"eval_runtime": 14.5,
|
| 2464 |
+
"eval_samples_per_second": 28.414,
|
| 2465 |
+
"eval_steps_per_second": 3.586,
|
| 2466 |
+
"step": 1350
|
| 2467 |
+
},
|
| 2468 |
+
{
|
| 2469 |
+
"epoch": 2.1318422601530314,
|
| 2470 |
+
"grad_norm": 2.377746343612671,
|
| 2471 |
+
"learning_rate": 1.067012561698319e-07,
|
| 2472 |
+
"logits/chosen": 12.19007396697998,
|
| 2473 |
+
"logits/rejected": 11.572507858276367,
|
| 2474 |
+
"logps/chosen": -4.9034600257873535,
|
| 2475 |
+
"logps/rejected": -9.414239883422852,
|
| 2476 |
+
"loss": 0.2174,
|
| 2477 |
+
"rewards/accuracies": 0.9375,
|
| 2478 |
+
"rewards/chosen": -7.355191230773926,
|
| 2479 |
+
"rewards/margins": 6.766169548034668,
|
| 2480 |
+
"rewards/rejected": -14.121360778808594,
|
| 2481 |
+
"step": 1360
|
| 2482 |
+
},
|
| 2483 |
+
{
|
| 2484 |
+
"epoch": 2.1475377673141063,
|
| 2485 |
+
"grad_norm": 3.618283748626709,
|
| 2486 |
+
"learning_rate": 9.209358300585474e-08,
|
| 2487 |
+
"logits/chosen": 12.282819747924805,
|
| 2488 |
+
"logits/rejected": 12.010876655578613,
|
| 2489 |
+
"logps/chosen": -4.678290367126465,
|
| 2490 |
+
"logps/rejected": -8.606224060058594,
|
| 2491 |
+
"loss": 0.2393,
|
| 2492 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 2493 |
+
"rewards/chosen": -7.0174360275268555,
|
| 2494 |
+
"rewards/margins": 5.89190149307251,
|
| 2495 |
+
"rewards/rejected": -12.909337043762207,
|
| 2496 |
+
"step": 1370
|
| 2497 |
+
},
|
| 2498 |
+
{
|
| 2499 |
+
"epoch": 2.1632332744751817,
|
| 2500 |
+
"grad_norm": 2.742673873901367,
|
| 2501 |
+
"learning_rate": 7.854209717842231e-08,
|
| 2502 |
+
"logits/chosen": 12.202973365783691,
|
| 2503 |
+
"logits/rejected": 11.41163444519043,
|
| 2504 |
+
"logps/chosen": -4.9238996505737305,
|
| 2505 |
+
"logps/rejected": -9.56473445892334,
|
| 2506 |
+
"loss": 0.1748,
|
| 2507 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 2508 |
+
"rewards/chosen": -7.3858489990234375,
|
| 2509 |
+
"rewards/margins": 6.9612531661987305,
|
| 2510 |
+
"rewards/rejected": -14.3471040725708,
|
| 2511 |
+
"step": 1380
|
| 2512 |
+
},
|
| 2513 |
+
{
|
| 2514 |
+
"epoch": 2.1789287816362566,
|
| 2515 |
+
"grad_norm": 2.435816764831543,
|
| 2516 |
+
"learning_rate": 6.605274281709929e-08,
|
| 2517 |
+
"logits/chosen": 12.051380157470703,
|
| 2518 |
+
"logits/rejected": 10.81779670715332,
|
| 2519 |
+
"logps/chosen": -4.7611918449401855,
|
| 2520 |
+
"logps/rejected": -8.955581665039062,
|
| 2521 |
+
"loss": 0.2023,
|
| 2522 |
+
"rewards/accuracies": 0.949999988079071,
|
| 2523 |
+
"rewards/chosen": -7.141787528991699,
|
| 2524 |
+
"rewards/margins": 6.291583061218262,
|
| 2525 |
+
"rewards/rejected": -13.433370590209961,
|
| 2526 |
+
"step": 1390
|
| 2527 |
+
},
|
| 2528 |
+
{
|
| 2529 |
+
"epoch": 2.194624288797332,
|
| 2530 |
+
"grad_norm": 3.3359460830688477,
|
| 2531 |
+
"learning_rate": 5.463099816548578e-08,
|
| 2532 |
+
"logits/chosen": 12.5380220413208,
|
| 2533 |
+
"logits/rejected": 11.374457359313965,
|
| 2534 |
+
"logps/chosen": -4.384824275970459,
|
| 2535 |
+
"logps/rejected": -9.00879192352295,
|
| 2536 |
+
"loss": 0.1709,
|
| 2537 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 2538 |
+
"rewards/chosen": -6.577237129211426,
|
| 2539 |
+
"rewards/margins": 6.935951232910156,
|
| 2540 |
+
"rewards/rejected": -13.513188362121582,
|
| 2541 |
+
"step": 1400
|
| 2542 |
+
},
|
| 2543 |
+
{
|
| 2544 |
+
"epoch": 2.194624288797332,
|
| 2545 |
+
"eval_logits/chosen": 12.820469856262207,
|
| 2546 |
+
"eval_logits/rejected": 11.797761917114258,
|
| 2547 |
+
"eval_logps/chosen": -4.198721408843994,
|
| 2548 |
+
"eval_logps/rejected": -8.741405487060547,
|
| 2549 |
+
"eval_loss": 0.23232795298099518,
|
| 2550 |
+
"eval_rewards/accuracies": 1.0,
|
| 2551 |
+
"eval_rewards/chosen": -6.29808235168457,
|
| 2552 |
+
"eval_rewards/margins": 6.81402587890625,
|
| 2553 |
+
"eval_rewards/rejected": -13.112107276916504,
|
| 2554 |
+
"eval_runtime": 14.5059,
|
| 2555 |
+
"eval_samples_per_second": 28.402,
|
| 2556 |
+
"eval_steps_per_second": 3.585,
|
| 2557 |
+
"step": 1400
|
| 2558 |
+
}
|
| 2559 |
+
],
|
| 2560 |
+
"logging_steps": 10,
|
| 2561 |
+
"max_steps": 1500,
|
| 2562 |
+
"num_input_tokens_seen": 0,
|
| 2563 |
+
"num_train_epochs": 3,
|
| 2564 |
+
"save_steps": 50,
|
| 2565 |
+
"stateful_callbacks": {
|
| 2566 |
+
"TrainerControl": {
|
| 2567 |
+
"args": {
|
| 2568 |
+
"should_epoch_stop": false,
|
| 2569 |
+
"should_evaluate": false,
|
| 2570 |
+
"should_log": false,
|
| 2571 |
+
"should_save": true,
|
| 2572 |
+
"should_training_stop": false
|
| 2573 |
+
},
|
| 2574 |
+
"attributes": {}
|
| 2575 |
+
}
|
| 2576 |
+
},
|
| 2577 |
+
"total_flos": 3.388940081585717e+18,
|
| 2578 |
+
"train_batch_size": 1,
|
| 2579 |
+
"trial_name": null,
|
| 2580 |
+
"trial_params": null
|
| 2581 |
+
}
|
checkpoint-1400/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9b2203cea5a5bcb79dc529b64d6033ba414f923d9d5e3c378f433027100e50b9
|
| 3 |
+
size 7224
|
checkpoint-1400/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)
|