Training in progress, step 1200, checkpoint
Browse files- checkpoint-1200/README.md +202 -0
- checkpoint-1200/adapter_config.json +34 -0
- checkpoint-1200/adapter_model.safetensors +3 -0
- checkpoint-1200/global_step1199/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt +3 -0
- checkpoint-1200/global_step1199/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt +3 -0
- checkpoint-1200/global_step1199/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt +3 -0
- checkpoint-1200/global_step1199/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt +3 -0
- checkpoint-1200/global_step1199/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt +3 -0
- checkpoint-1200/global_step1199/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt +3 -0
- checkpoint-1200/global_step1199/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt +3 -0
- checkpoint-1200/global_step1199/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt +3 -0
- checkpoint-1200/global_step1199/mp_rank_00_model_states.pt +3 -0
- checkpoint-1200/latest +1 -0
- checkpoint-1200/rng_state_0.pth +3 -0
- checkpoint-1200/rng_state_1.pth +3 -0
- checkpoint-1200/rng_state_2.pth +3 -0
- checkpoint-1200/rng_state_3.pth +3 -0
- checkpoint-1200/rng_state_4.pth +3 -0
- checkpoint-1200/rng_state_5.pth +3 -0
- checkpoint-1200/rng_state_6.pth +3 -0
- checkpoint-1200/rng_state_7.pth +3 -0
- checkpoint-1200/scheduler.pt +3 -0
- checkpoint-1200/special_tokens_map.json +30 -0
- checkpoint-1200/tokenizer.json +0 -0
- checkpoint-1200/tokenizer_config.json +133 -0
- checkpoint-1200/trainer_state.json +2217 -0
- checkpoint-1200/training_args.bin +3 -0
- checkpoint-1200/zero_to_fp32.py +674 -0
checkpoint-1200/README.md
ADDED
|
@@ -0,0 +1,202 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
base_model: microsoft/Phi-3-mini-4k-instruct
|
| 3 |
+
library_name: peft
|
| 4 |
+
---
|
| 5 |
+
|
| 6 |
+
# Model Card for Model ID
|
| 7 |
+
|
| 8 |
+
<!-- Provide a quick summary of what the model is/does. -->
|
| 9 |
+
|
| 10 |
+
|
| 11 |
+
|
| 12 |
+
## Model Details
|
| 13 |
+
|
| 14 |
+
### Model Description
|
| 15 |
+
|
| 16 |
+
<!-- Provide a longer summary of what this model is. -->
|
| 17 |
+
|
| 18 |
+
|
| 19 |
+
|
| 20 |
+
- **Developed by:** [More Information Needed]
|
| 21 |
+
- **Funded by [optional]:** [More Information Needed]
|
| 22 |
+
- **Shared by [optional]:** [More Information Needed]
|
| 23 |
+
- **Model type:** [More Information Needed]
|
| 24 |
+
- **Language(s) (NLP):** [More Information Needed]
|
| 25 |
+
- **License:** [More Information Needed]
|
| 26 |
+
- **Finetuned from model [optional]:** [More Information Needed]
|
| 27 |
+
|
| 28 |
+
### Model Sources [optional]
|
| 29 |
+
|
| 30 |
+
<!-- Provide the basic links for the model. -->
|
| 31 |
+
|
| 32 |
+
- **Repository:** [More Information Needed]
|
| 33 |
+
- **Paper [optional]:** [More Information Needed]
|
| 34 |
+
- **Demo [optional]:** [More Information Needed]
|
| 35 |
+
|
| 36 |
+
## Uses
|
| 37 |
+
|
| 38 |
+
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
|
| 39 |
+
|
| 40 |
+
### Direct Use
|
| 41 |
+
|
| 42 |
+
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
|
| 43 |
+
|
| 44 |
+
[More Information Needed]
|
| 45 |
+
|
| 46 |
+
### Downstream Use [optional]
|
| 47 |
+
|
| 48 |
+
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
|
| 49 |
+
|
| 50 |
+
[More Information Needed]
|
| 51 |
+
|
| 52 |
+
### Out-of-Scope Use
|
| 53 |
+
|
| 54 |
+
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
|
| 55 |
+
|
| 56 |
+
[More Information Needed]
|
| 57 |
+
|
| 58 |
+
## Bias, Risks, and Limitations
|
| 59 |
+
|
| 60 |
+
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
|
| 61 |
+
|
| 62 |
+
[More Information Needed]
|
| 63 |
+
|
| 64 |
+
### Recommendations
|
| 65 |
+
|
| 66 |
+
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
|
| 67 |
+
|
| 68 |
+
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
|
| 69 |
+
|
| 70 |
+
## How to Get Started with the Model
|
| 71 |
+
|
| 72 |
+
Use the code below to get started with the model.
|
| 73 |
+
|
| 74 |
+
[More Information Needed]
|
| 75 |
+
|
| 76 |
+
## Training Details
|
| 77 |
+
|
| 78 |
+
### Training Data
|
| 79 |
+
|
| 80 |
+
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
|
| 81 |
+
|
| 82 |
+
[More Information Needed]
|
| 83 |
+
|
| 84 |
+
### Training Procedure
|
| 85 |
+
|
| 86 |
+
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
|
| 87 |
+
|
| 88 |
+
#### Preprocessing [optional]
|
| 89 |
+
|
| 90 |
+
[More Information Needed]
|
| 91 |
+
|
| 92 |
+
|
| 93 |
+
#### Training Hyperparameters
|
| 94 |
+
|
| 95 |
+
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
|
| 96 |
+
|
| 97 |
+
#### Speeds, Sizes, Times [optional]
|
| 98 |
+
|
| 99 |
+
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
|
| 100 |
+
|
| 101 |
+
[More Information Needed]
|
| 102 |
+
|
| 103 |
+
## Evaluation
|
| 104 |
+
|
| 105 |
+
<!-- This section describes the evaluation protocols and provides the results. -->
|
| 106 |
+
|
| 107 |
+
### Testing Data, Factors & Metrics
|
| 108 |
+
|
| 109 |
+
#### Testing Data
|
| 110 |
+
|
| 111 |
+
<!-- This should link to a Dataset Card if possible. -->
|
| 112 |
+
|
| 113 |
+
[More Information Needed]
|
| 114 |
+
|
| 115 |
+
#### Factors
|
| 116 |
+
|
| 117 |
+
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
|
| 118 |
+
|
| 119 |
+
[More Information Needed]
|
| 120 |
+
|
| 121 |
+
#### Metrics
|
| 122 |
+
|
| 123 |
+
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
|
| 124 |
+
|
| 125 |
+
[More Information Needed]
|
| 126 |
+
|
| 127 |
+
### Results
|
| 128 |
+
|
| 129 |
+
[More Information Needed]
|
| 130 |
+
|
| 131 |
+
#### Summary
|
| 132 |
+
|
| 133 |
+
|
| 134 |
+
|
| 135 |
+
## Model Examination [optional]
|
| 136 |
+
|
| 137 |
+
<!-- Relevant interpretability work for the model goes here -->
|
| 138 |
+
|
| 139 |
+
[More Information Needed]
|
| 140 |
+
|
| 141 |
+
## Environmental Impact
|
| 142 |
+
|
| 143 |
+
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
|
| 144 |
+
|
| 145 |
+
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
|
| 146 |
+
|
| 147 |
+
- **Hardware Type:** [More Information Needed]
|
| 148 |
+
- **Hours used:** [More Information Needed]
|
| 149 |
+
- **Cloud Provider:** [More Information Needed]
|
| 150 |
+
- **Compute Region:** [More Information Needed]
|
| 151 |
+
- **Carbon Emitted:** [More Information Needed]
|
| 152 |
+
|
| 153 |
+
## Technical Specifications [optional]
|
| 154 |
+
|
| 155 |
+
### Model Architecture and Objective
|
| 156 |
+
|
| 157 |
+
[More Information Needed]
|
| 158 |
+
|
| 159 |
+
### Compute Infrastructure
|
| 160 |
+
|
| 161 |
+
[More Information Needed]
|
| 162 |
+
|
| 163 |
+
#### Hardware
|
| 164 |
+
|
| 165 |
+
[More Information Needed]
|
| 166 |
+
|
| 167 |
+
#### Software
|
| 168 |
+
|
| 169 |
+
[More Information Needed]
|
| 170 |
+
|
| 171 |
+
## Citation [optional]
|
| 172 |
+
|
| 173 |
+
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
|
| 174 |
+
|
| 175 |
+
**BibTeX:**
|
| 176 |
+
|
| 177 |
+
[More Information Needed]
|
| 178 |
+
|
| 179 |
+
**APA:**
|
| 180 |
+
|
| 181 |
+
[More Information Needed]
|
| 182 |
+
|
| 183 |
+
## Glossary [optional]
|
| 184 |
+
|
| 185 |
+
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
|
| 186 |
+
|
| 187 |
+
[More Information Needed]
|
| 188 |
+
|
| 189 |
+
## More Information [optional]
|
| 190 |
+
|
| 191 |
+
[More Information Needed]
|
| 192 |
+
|
| 193 |
+
## Model Card Authors [optional]
|
| 194 |
+
|
| 195 |
+
[More Information Needed]
|
| 196 |
+
|
| 197 |
+
## Model Card Contact
|
| 198 |
+
|
| 199 |
+
[More Information Needed]
|
| 200 |
+
### Framework versions
|
| 201 |
+
|
| 202 |
+
- PEFT 0.14.0
|
checkpoint-1200/adapter_config.json
ADDED
|
@@ -0,0 +1,34 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"alpha_pattern": {},
|
| 3 |
+
"auto_mapping": null,
|
| 4 |
+
"base_model_name_or_path": "microsoft/Phi-3-mini-4k-instruct",
|
| 5 |
+
"bias": "none",
|
| 6 |
+
"eva_config": null,
|
| 7 |
+
"exclude_modules": null,
|
| 8 |
+
"fan_in_fan_out": false,
|
| 9 |
+
"inference_mode": true,
|
| 10 |
+
"init_lora_weights": true,
|
| 11 |
+
"layer_replication": null,
|
| 12 |
+
"layers_pattern": null,
|
| 13 |
+
"layers_to_transform": null,
|
| 14 |
+
"loftq_config": {},
|
| 15 |
+
"lora_alpha": 16,
|
| 16 |
+
"lora_bias": false,
|
| 17 |
+
"lora_dropout": 0.0,
|
| 18 |
+
"megatron_config": null,
|
| 19 |
+
"megatron_core": "megatron.core",
|
| 20 |
+
"modules_to_save": null,
|
| 21 |
+
"peft_type": "LORA",
|
| 22 |
+
"r": 8,
|
| 23 |
+
"rank_pattern": {},
|
| 24 |
+
"revision": null,
|
| 25 |
+
"target_modules": [
|
| 26 |
+
"qkv_proj",
|
| 27 |
+
"down_proj",
|
| 28 |
+
"gate_up_proj",
|
| 29 |
+
"o_proj"
|
| 30 |
+
],
|
| 31 |
+
"task_type": "CAUSAL_LM",
|
| 32 |
+
"use_dora": false,
|
| 33 |
+
"use_rslora": false
|
| 34 |
+
}
|
checkpoint-1200/adapter_model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f833b957652935b9372a7ea25a47b1ccad67b7610e560bb5ee3c7e3c8964a5dc
|
| 3 |
+
size 25200088
|
checkpoint-1200/global_step1199/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:d305e148ca407edbc8bf0c3c9302b73ce9a131dcfa4b513c8163d46f8995ee54
|
| 3 |
+
size 18881328
|
checkpoint-1200/global_step1199/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 sha256:8debbc266ea10ffe488bf69be1d0ee8ef3a3e8ae30952b0f70139ad488ae5a96
|
| 3 |
+
size 18881328
|
checkpoint-1200/global_step1199/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 sha256:72403a6c8815fc1b8728907aebb279fe715411aef6a8cb0ff947953cd8ac0f75
|
| 3 |
+
size 18881328
|
checkpoint-1200/global_step1199/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 sha256:9e3b925f5279416186ec013c6a28ce9830f3500a6c4114e5e049d6190a108dd6
|
| 3 |
+
size 18881392
|
checkpoint-1200/global_step1199/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:f443e00aabb50350f6ca1fdea6e2fc9e98e9a34c9e208484569825cf5d226495
|
| 3 |
+
size 18881392
|
checkpoint-1200/global_step1199/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 sha256:5a58b9d429f26f4e63d98d5e7e0f9b5f33d2a96e3daf31c40e2ea76f0b4dafb0
|
| 3 |
+
size 18881392
|
checkpoint-1200/global_step1199/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 sha256:9e24ebd27b09e3e6f952936138b65259a6f87ddb0538af0bd51e7c00d68a6435
|
| 3 |
+
size 18881392
|
checkpoint-1200/global_step1199/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 sha256:6ebcf5c306c8f19487377c401d29eb1b24239791877dd678fd945298310f627c
|
| 3 |
+
size 18881392
|
checkpoint-1200/global_step1199/mp_rank_00_model_states.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:95264200eb57bdb89acc3cc11ec97f8c4dde97d60e697557588ff4e62dd3f592
|
| 3 |
+
size 25379244
|
checkpoint-1200/latest
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
global_step1199
|
checkpoint-1200/rng_state_0.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:479e628a41471621ed71e5888d4bc3eecd148e78c535bc6314d7d47b1c46d139
|
| 3 |
+
size 15984
|
checkpoint-1200/rng_state_1.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:977d4030d8e6485cb5e6f8932d356d077b621fc5e3a3a17c1cd7300e3c3aa378
|
| 3 |
+
size 15984
|
checkpoint-1200/rng_state_2.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f6cee236c0b6ea406dfd78dc5d74c5d225be0fa30781914f3736a59b5bf50da8
|
| 3 |
+
size 15984
|
checkpoint-1200/rng_state_3.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:151976a0e4efc5762247214e2e9304a5c082e878962f2b322e00eb30ef03cab3
|
| 3 |
+
size 15984
|
checkpoint-1200/rng_state_4.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ea64ac8295f29f6c548b206fafc96dea4fe6bf7d44797d6d0470afa88b072460
|
| 3 |
+
size 15984
|
checkpoint-1200/rng_state_5.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f61787141f0f1134f74a65e87d4d3ea74e81f6b9d697755ee30ef0b88b18ddba
|
| 3 |
+
size 15984
|
checkpoint-1200/rng_state_6.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:787193a355e21da1f4d95b759a157a105a86b3062286f0d9df1eb7fd2c5cd54c
|
| 3 |
+
size 15984
|
checkpoint-1200/rng_state_7.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c8b1b9e4e9a20f686ee92a48f0a8089eb687a9e2a7b8be07f71fdf84e2fcf1a4
|
| 3 |
+
size 15984
|
checkpoint-1200/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:35b4cc7927ca0f1b9b45cd00f72746408c82d953cb952c75e7569243d9fa3f0c
|
| 3 |
+
size 1064
|
checkpoint-1200/special_tokens_map.json
ADDED
|
@@ -0,0 +1,30 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"content": "<s>",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": false,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"eos_token": {
|
| 10 |
+
"content": "<|end|>",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": false,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"pad_token": {
|
| 17 |
+
"content": "<|endoftext|>",
|
| 18 |
+
"lstrip": false,
|
| 19 |
+
"normalized": false,
|
| 20 |
+
"rstrip": false,
|
| 21 |
+
"single_word": false
|
| 22 |
+
},
|
| 23 |
+
"unk_token": {
|
| 24 |
+
"content": "<unk>",
|
| 25 |
+
"lstrip": false,
|
| 26 |
+
"normalized": false,
|
| 27 |
+
"rstrip": false,
|
| 28 |
+
"single_word": false
|
| 29 |
+
}
|
| 30 |
+
}
|
checkpoint-1200/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
checkpoint-1200/tokenizer_config.json
ADDED
|
@@ -0,0 +1,133 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_bos_token": false,
|
| 3 |
+
"add_eos_token": false,
|
| 4 |
+
"add_prefix_space": null,
|
| 5 |
+
"added_tokens_decoder": {
|
| 6 |
+
"0": {
|
| 7 |
+
"content": "<unk>",
|
| 8 |
+
"lstrip": false,
|
| 9 |
+
"normalized": false,
|
| 10 |
+
"rstrip": false,
|
| 11 |
+
"single_word": false,
|
| 12 |
+
"special": true
|
| 13 |
+
},
|
| 14 |
+
"1": {
|
| 15 |
+
"content": "<s>",
|
| 16 |
+
"lstrip": false,
|
| 17 |
+
"normalized": false,
|
| 18 |
+
"rstrip": false,
|
| 19 |
+
"single_word": false,
|
| 20 |
+
"special": true
|
| 21 |
+
},
|
| 22 |
+
"2": {
|
| 23 |
+
"content": "</s>",
|
| 24 |
+
"lstrip": false,
|
| 25 |
+
"normalized": false,
|
| 26 |
+
"rstrip": true,
|
| 27 |
+
"single_word": false,
|
| 28 |
+
"special": false
|
| 29 |
+
},
|
| 30 |
+
"32000": {
|
| 31 |
+
"content": "<|endoftext|>",
|
| 32 |
+
"lstrip": false,
|
| 33 |
+
"normalized": false,
|
| 34 |
+
"rstrip": false,
|
| 35 |
+
"single_word": false,
|
| 36 |
+
"special": true
|
| 37 |
+
},
|
| 38 |
+
"32001": {
|
| 39 |
+
"content": "<|assistant|>",
|
| 40 |
+
"lstrip": false,
|
| 41 |
+
"normalized": false,
|
| 42 |
+
"rstrip": true,
|
| 43 |
+
"single_word": false,
|
| 44 |
+
"special": true
|
| 45 |
+
},
|
| 46 |
+
"32002": {
|
| 47 |
+
"content": "<|placeholder1|>",
|
| 48 |
+
"lstrip": false,
|
| 49 |
+
"normalized": false,
|
| 50 |
+
"rstrip": true,
|
| 51 |
+
"single_word": false,
|
| 52 |
+
"special": true
|
| 53 |
+
},
|
| 54 |
+
"32003": {
|
| 55 |
+
"content": "<|placeholder2|>",
|
| 56 |
+
"lstrip": false,
|
| 57 |
+
"normalized": false,
|
| 58 |
+
"rstrip": true,
|
| 59 |
+
"single_word": false,
|
| 60 |
+
"special": true
|
| 61 |
+
},
|
| 62 |
+
"32004": {
|
| 63 |
+
"content": "<|placeholder3|>",
|
| 64 |
+
"lstrip": false,
|
| 65 |
+
"normalized": false,
|
| 66 |
+
"rstrip": true,
|
| 67 |
+
"single_word": false,
|
| 68 |
+
"special": true
|
| 69 |
+
},
|
| 70 |
+
"32005": {
|
| 71 |
+
"content": "<|placeholder4|>",
|
| 72 |
+
"lstrip": false,
|
| 73 |
+
"normalized": false,
|
| 74 |
+
"rstrip": true,
|
| 75 |
+
"single_word": false,
|
| 76 |
+
"special": true
|
| 77 |
+
},
|
| 78 |
+
"32006": {
|
| 79 |
+
"content": "<|system|>",
|
| 80 |
+
"lstrip": false,
|
| 81 |
+
"normalized": false,
|
| 82 |
+
"rstrip": true,
|
| 83 |
+
"single_word": false,
|
| 84 |
+
"special": true
|
| 85 |
+
},
|
| 86 |
+
"32007": {
|
| 87 |
+
"content": "<|end|>",
|
| 88 |
+
"lstrip": false,
|
| 89 |
+
"normalized": false,
|
| 90 |
+
"rstrip": false,
|
| 91 |
+
"single_word": false,
|
| 92 |
+
"special": true
|
| 93 |
+
},
|
| 94 |
+
"32008": {
|
| 95 |
+
"content": "<|placeholder5|>",
|
| 96 |
+
"lstrip": false,
|
| 97 |
+
"normalized": false,
|
| 98 |
+
"rstrip": true,
|
| 99 |
+
"single_word": false,
|
| 100 |
+
"special": true
|
| 101 |
+
},
|
| 102 |
+
"32009": {
|
| 103 |
+
"content": "<|placeholder6|>",
|
| 104 |
+
"lstrip": false,
|
| 105 |
+
"normalized": false,
|
| 106 |
+
"rstrip": true,
|
| 107 |
+
"single_word": false,
|
| 108 |
+
"special": true
|
| 109 |
+
},
|
| 110 |
+
"32010": {
|
| 111 |
+
"content": "<|user|>",
|
| 112 |
+
"lstrip": false,
|
| 113 |
+
"normalized": false,
|
| 114 |
+
"rstrip": true,
|
| 115 |
+
"single_word": false,
|
| 116 |
+
"special": true
|
| 117 |
+
}
|
| 118 |
+
},
|
| 119 |
+
"bos_token": "<s>",
|
| 120 |
+
"chat_template": "{% set system_message = 'You are a helpful AI assistant.' %}{% if messages[0]['role'] == 'system' %}{% set system_message = messages[0]['content'] %}{% endif %}{% if system_message is defined %}{{ '<s>' + '<|system|>\n' + system_message + '<|end|>\n' }}{% endif %}{% for message in messages %}{% set content = message['content'] %}{% if message['role'] == 'user' %}{{ '<|user|>\n' + content + '<|end|>\n<|assistant|>\n' }}{% elif message['role'] == 'assistant' %}{{ content + '<|end|>' + '\n' }}{% endif %}{% endfor %}",
|
| 121 |
+
"clean_up_tokenization_spaces": false,
|
| 122 |
+
"eos_token": "<|end|>",
|
| 123 |
+
"extra_special_tokens": {},
|
| 124 |
+
"legacy": false,
|
| 125 |
+
"model_max_length": 4096,
|
| 126 |
+
"pad_token": "<|endoftext|>",
|
| 127 |
+
"padding_side": "right",
|
| 128 |
+
"sp_model_kwargs": {},
|
| 129 |
+
"split_special_tokens": false,
|
| 130 |
+
"tokenizer_class": "LlamaTokenizer",
|
| 131 |
+
"unk_token": "<unk>",
|
| 132 |
+
"use_default_system_prompt": false
|
| 133 |
+
}
|
checkpoint-1200/trainer_state.json
ADDED
|
@@ -0,0 +1,2217 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 1.882087502452423,
|
| 5 |
+
"eval_steps": 50,
|
| 6 |
+
"global_step": 1200,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.015695507161075144,
|
| 13 |
+
"grad_norm": 0.07061820477247238,
|
| 14 |
+
"learning_rate": 4.999451708687114e-06,
|
| 15 |
+
"logits/chosen": 15.001790046691895,
|
| 16 |
+
"logits/rejected": 14.624488830566406,
|
| 17 |
+
"logps/chosen": -0.33085882663726807,
|
| 18 |
+
"logps/rejected": -0.24924471974372864,
|
| 19 |
+
"loss": 1.0519,
|
| 20 |
+
"rewards/accuracies": 0.23749999701976776,
|
| 21 |
+
"rewards/chosen": -0.4962882399559021,
|
| 22 |
+
"rewards/margins": -0.12242116779088974,
|
| 23 |
+
"rewards/rejected": -0.37386709451675415,
|
| 24 |
+
"step": 10
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.03139101432215029,
|
| 28 |
+
"grad_norm": 0.060370542109012604,
|
| 29 |
+
"learning_rate": 4.997807075247147e-06,
|
| 30 |
+
"logits/chosen": 15.272351264953613,
|
| 31 |
+
"logits/rejected": 14.801017761230469,
|
| 32 |
+
"logps/chosen": -0.3379867672920227,
|
| 33 |
+
"logps/rejected": -0.24759705364704132,
|
| 34 |
+
"loss": 1.0494,
|
| 35 |
+
"rewards/accuracies": 0.15000000596046448,
|
| 36 |
+
"rewards/chosen": -0.5069801211357117,
|
| 37 |
+
"rewards/margins": -0.13558456301689148,
|
| 38 |
+
"rewards/rejected": -0.3713955581188202,
|
| 39 |
+
"step": 20
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.047086521483225424,
|
| 43 |
+
"grad_norm": 0.050332240760326385,
|
| 44 |
+
"learning_rate": 4.9950668210706795e-06,
|
| 45 |
+
"logits/chosen": 15.569372177124023,
|
| 46 |
+
"logits/rejected": 15.249380111694336,
|
| 47 |
+
"logps/chosen": -0.3347483277320862,
|
| 48 |
+
"logps/rejected": -0.275604784488678,
|
| 49 |
+
"loss": 1.0375,
|
| 50 |
+
"rewards/accuracies": 0.21250000596046448,
|
| 51 |
+
"rewards/chosen": -0.5021225214004517,
|
| 52 |
+
"rewards/margins": -0.08871528506278992,
|
| 53 |
+
"rewards/rejected": -0.41340717673301697,
|
| 54 |
+
"step": 30
|
| 55 |
+
},
|
| 56 |
+
{
|
| 57 |
+
"epoch": 0.06278202864430057,
|
| 58 |
+
"grad_norm": 0.054660990834236145,
|
| 59 |
+
"learning_rate": 4.9912321481237616e-06,
|
| 60 |
+
"logits/chosen": 15.344067573547363,
|
| 61 |
+
"logits/rejected": 14.881669998168945,
|
| 62 |
+
"logps/chosen": -0.3336474299430847,
|
| 63 |
+
"logps/rejected": -0.27252259850502014,
|
| 64 |
+
"loss": 1.0375,
|
| 65 |
+
"rewards/accuracies": 0.20000000298023224,
|
| 66 |
+
"rewards/chosen": -0.5004712343215942,
|
| 67 |
+
"rewards/margins": -0.09168727695941925,
|
| 68 |
+
"rewards/rejected": -0.40878385305404663,
|
| 69 |
+
"step": 40
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 0.07847753580537571,
|
| 73 |
+
"grad_norm": 0.060405269265174866,
|
| 74 |
+
"learning_rate": 4.986304738420684e-06,
|
| 75 |
+
"logits/chosen": 15.5701904296875,
|
| 76 |
+
"logits/rejected": 15.60998821258545,
|
| 77 |
+
"logps/chosen": -0.3075069785118103,
|
| 78 |
+
"logps/rejected": -0.24850216507911682,
|
| 79 |
+
"loss": 1.0341,
|
| 80 |
+
"rewards/accuracies": 0.20000000298023224,
|
| 81 |
+
"rewards/chosen": -0.46126049757003784,
|
| 82 |
+
"rewards/margins": -0.08850721269845963,
|
| 83 |
+
"rewards/rejected": -0.3727532923221588,
|
| 84 |
+
"step": 50
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 0.07847753580537571,
|
| 88 |
+
"eval_logits/chosen": 15.719181060791016,
|
| 89 |
+
"eval_logits/rejected": 15.20205307006836,
|
| 90 |
+
"eval_logps/chosen": -0.32490479946136475,
|
| 91 |
+
"eval_logps/rejected": -0.26058924198150635,
|
| 92 |
+
"eval_loss": 1.0295902490615845,
|
| 93 |
+
"eval_rewards/accuracies": 0.26923078298568726,
|
| 94 |
+
"eval_rewards/chosen": -0.48735716938972473,
|
| 95 |
+
"eval_rewards/margins": -0.09647335112094879,
|
| 96 |
+
"eval_rewards/rejected": -0.39088380336761475,
|
| 97 |
+
"eval_runtime": 14.6102,
|
| 98 |
+
"eval_samples_per_second": 28.199,
|
| 99 |
+
"eval_steps_per_second": 3.559,
|
| 100 |
+
"step": 50
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.09417304296645085,
|
| 104 |
+
"grad_norm": 0.06511708348989487,
|
| 105 |
+
"learning_rate": 4.980286753286196e-06,
|
| 106 |
+
"logits/chosen": 15.571348190307617,
|
| 107 |
+
"logits/rejected": 15.418828010559082,
|
| 108 |
+
"logps/chosen": -0.3217025101184845,
|
| 109 |
+
"logps/rejected": -0.2813836932182312,
|
| 110 |
+
"loss": 1.0385,
|
| 111 |
+
"rewards/accuracies": 0.23749999701976776,
|
| 112 |
+
"rewards/chosen": -0.48255378007888794,
|
| 113 |
+
"rewards/margins": -0.060478221625089645,
|
| 114 |
+
"rewards/rejected": -0.4220755696296692,
|
| 115 |
+
"step": 60
|
| 116 |
+
},
|
| 117 |
+
{
|
| 118 |
+
"epoch": 0.109868550127526,
|
| 119 |
+
"grad_norm": 0.06922140717506409,
|
| 120 |
+
"learning_rate": 4.973180832407471e-06,
|
| 121 |
+
"logits/chosen": 15.926614761352539,
|
| 122 |
+
"logits/rejected": 15.90565299987793,
|
| 123 |
+
"logps/chosen": -0.34742942452430725,
|
| 124 |
+
"logps/rejected": -0.26853513717651367,
|
| 125 |
+
"loss": 1.0417,
|
| 126 |
+
"rewards/accuracies": 0.21250000596046448,
|
| 127 |
+
"rewards/chosen": -0.5211440324783325,
|
| 128 |
+
"rewards/margins": -0.11834144592285156,
|
| 129 |
+
"rewards/rejected": -0.40280264616012573,
|
| 130 |
+
"step": 70
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 0.12556405728860115,
|
| 134 |
+
"grad_norm": 0.07853339612483978,
|
| 135 |
+
"learning_rate": 4.964990092676263e-06,
|
| 136 |
+
"logits/chosen": 15.876957893371582,
|
| 137 |
+
"logits/rejected": 15.636863708496094,
|
| 138 |
+
"logps/chosen": -0.34999752044677734,
|
| 139 |
+
"logps/rejected": -0.2652502655982971,
|
| 140 |
+
"loss": 1.0255,
|
| 141 |
+
"rewards/accuracies": 0.1875,
|
| 142 |
+
"rewards/chosen": -0.524996280670166,
|
| 143 |
+
"rewards/margins": -0.12712089717388153,
|
| 144 |
+
"rewards/rejected": -0.3978753685951233,
|
| 145 |
+
"step": 80
|
| 146 |
+
},
|
| 147 |
+
{
|
| 148 |
+
"epoch": 0.14125956444967627,
|
| 149 |
+
"grad_norm": 0.07262148708105087,
|
| 150 |
+
"learning_rate": 4.9557181268217225e-06,
|
| 151 |
+
"logits/chosen": 16.037456512451172,
|
| 152 |
+
"logits/rejected": 15.742408752441406,
|
| 153 |
+
"logps/chosen": -0.32482510805130005,
|
| 154 |
+
"logps/rejected": -0.2520079016685486,
|
| 155 |
+
"loss": 1.0297,
|
| 156 |
+
"rewards/accuracies": 0.21250000596046448,
|
| 157 |
+
"rewards/chosen": -0.48723769187927246,
|
| 158 |
+
"rewards/margins": -0.1092257872223854,
|
| 159 |
+
"rewards/rejected": -0.3780118525028229,
|
| 160 |
+
"step": 90
|
| 161 |
+
},
|
| 162 |
+
{
|
| 163 |
+
"epoch": 0.15695507161075142,
|
| 164 |
+
"grad_norm": 0.08407289534807205,
|
| 165 |
+
"learning_rate": 4.9453690018345144e-06,
|
| 166 |
+
"logits/chosen": 16.34904670715332,
|
| 167 |
+
"logits/rejected": 16.30767059326172,
|
| 168 |
+
"logps/chosen": -0.3302846848964691,
|
| 169 |
+
"logps/rejected": -0.2799247205257416,
|
| 170 |
+
"loss": 1.0285,
|
| 171 |
+
"rewards/accuracies": 0.2750000059604645,
|
| 172 |
+
"rewards/chosen": -0.4954269826412201,
|
| 173 |
+
"rewards/margins": -0.07553993165493011,
|
| 174 |
+
"rewards/rejected": -0.41988706588745117,
|
| 175 |
+
"step": 100
|
| 176 |
+
},
|
| 177 |
+
{
|
| 178 |
+
"epoch": 0.15695507161075142,
|
| 179 |
+
"eval_logits/chosen": 16.43848419189453,
|
| 180 |
+
"eval_logits/rejected": 15.978095054626465,
|
| 181 |
+
"eval_logps/chosen": -0.3256901502609253,
|
| 182 |
+
"eval_logps/rejected": -0.2794351279735565,
|
| 183 |
+
"eval_loss": 1.0174708366394043,
|
| 184 |
+
"eval_rewards/accuracies": 0.2884615361690521,
|
| 185 |
+
"eval_rewards/chosen": -0.4885352551937103,
|
| 186 |
+
"eval_rewards/margins": -0.06938254088163376,
|
| 187 |
+
"eval_rewards/rejected": -0.41915270686149597,
|
| 188 |
+
"eval_runtime": 14.607,
|
| 189 |
+
"eval_samples_per_second": 28.206,
|
| 190 |
+
"eval_steps_per_second": 3.56,
|
| 191 |
+
"step": 100
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.17265057877182657,
|
| 195 |
+
"grad_norm": 0.07448805868625641,
|
| 196 |
+
"learning_rate": 4.933947257182901e-06,
|
| 197 |
+
"logits/chosen": 16.681476593017578,
|
| 198 |
+
"logits/rejected": 16.208908081054688,
|
| 199 |
+
"logps/chosen": -0.3410753309726715,
|
| 200 |
+
"logps/rejected": -0.28177398443222046,
|
| 201 |
+
"loss": 1.0364,
|
| 202 |
+
"rewards/accuracies": 0.2874999940395355,
|
| 203 |
+
"rewards/chosen": -0.5116130113601685,
|
| 204 |
+
"rewards/margins": -0.08895199000835419,
|
| 205 |
+
"rewards/rejected": -0.4226610064506531,
|
| 206 |
+
"step": 110
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.1883460859329017,
|
| 210 |
+
"grad_norm": 0.17266640067100525,
|
| 211 |
+
"learning_rate": 4.921457902821578e-06,
|
| 212 |
+
"logits/chosen": 16.73276710510254,
|
| 213 |
+
"logits/rejected": 16.46255111694336,
|
| 214 |
+
"logps/chosen": -0.30808666348457336,
|
| 215 |
+
"logps/rejected": -0.2663067877292633,
|
| 216 |
+
"loss": 1.0242,
|
| 217 |
+
"rewards/accuracies": 0.23749999701976776,
|
| 218 |
+
"rewards/chosen": -0.46213001012802124,
|
| 219 |
+
"rewards/margins": -0.0626697838306427,
|
| 220 |
+
"rewards/rejected": -0.39946022629737854,
|
| 221 |
+
"step": 120
|
| 222 |
+
},
|
| 223 |
+
{
|
| 224 |
+
"epoch": 0.20404159309397685,
|
| 225 |
+
"grad_norm": 0.1035998985171318,
|
| 226 |
+
"learning_rate": 4.907906416994146e-06,
|
| 227 |
+
"logits/chosen": 16.380239486694336,
|
| 228 |
+
"logits/rejected": 16.233612060546875,
|
| 229 |
+
"logps/chosen": -0.32250356674194336,
|
| 230 |
+
"logps/rejected": -0.27460020780563354,
|
| 231 |
+
"loss": 1.0162,
|
| 232 |
+
"rewards/accuracies": 0.2874999940395355,
|
| 233 |
+
"rewards/chosen": -0.48375529050827026,
|
| 234 |
+
"rewards/margins": -0.0718550831079483,
|
| 235 |
+
"rewards/rejected": -0.41190028190612793,
|
| 236 |
+
"step": 130
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 0.219737100255052,
|
| 240 |
+
"grad_norm": 0.0961030125617981,
|
| 241 |
+
"learning_rate": 4.893298743830168e-06,
|
| 242 |
+
"logits/chosen": 16.440303802490234,
|
| 243 |
+
"logits/rejected": 16.233903884887695,
|
| 244 |
+
"logps/chosen": -0.33421438932418823,
|
| 245 |
+
"logps/rejected": -0.2857271134853363,
|
| 246 |
+
"loss": 1.0008,
|
| 247 |
+
"rewards/accuracies": 0.3125,
|
| 248 |
+
"rewards/chosen": -0.5013214945793152,
|
| 249 |
+
"rewards/margins": -0.07273083180189133,
|
| 250 |
+
"rewards/rejected": -0.42859068512916565,
|
| 251 |
+
"step": 140
|
| 252 |
+
},
|
| 253 |
+
{
|
| 254 |
+
"epoch": 0.23543260741612712,
|
| 255 |
+
"grad_norm": 0.14799565076828003,
|
| 256 |
+
"learning_rate": 4.8776412907378845e-06,
|
| 257 |
+
"logits/chosen": 16.600738525390625,
|
| 258 |
+
"logits/rejected": 16.536447525024414,
|
| 259 |
+
"logps/chosen": -0.3243326246738434,
|
| 260 |
+
"logps/rejected": -0.3061785101890564,
|
| 261 |
+
"loss": 1.0085,
|
| 262 |
+
"rewards/accuracies": 0.3375000059604645,
|
| 263 |
+
"rewards/chosen": -0.4864989221096039,
|
| 264 |
+
"rewards/margins": -0.027231160551309586,
|
| 265 |
+
"rewards/rejected": -0.459267795085907,
|
| 266 |
+
"step": 150
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 0.23543260741612712,
|
| 270 |
+
"eval_logits/chosen": 16.8078556060791,
|
| 271 |
+
"eval_logits/rejected": 16.311922073364258,
|
| 272 |
+
"eval_logps/chosen": -0.326242595911026,
|
| 273 |
+
"eval_logps/rejected": -0.3398977816104889,
|
| 274 |
+
"eval_loss": 0.9791463017463684,
|
| 275 |
+
"eval_rewards/accuracies": 0.4038461446762085,
|
| 276 |
+
"eval_rewards/chosen": -0.4893638789653778,
|
| 277 |
+
"eval_rewards/margins": 0.020482787862420082,
|
| 278 |
+
"eval_rewards/rejected": -0.5098467469215393,
|
| 279 |
+
"eval_runtime": 14.6072,
|
| 280 |
+
"eval_samples_per_second": 28.205,
|
| 281 |
+
"eval_steps_per_second": 3.56,
|
| 282 |
+
"step": 150
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 0.2511281145772023,
|
| 286 |
+
"grad_norm": 0.11464422941207886,
|
| 287 |
+
"learning_rate": 4.860940925593703e-06,
|
| 288 |
+
"logits/chosen": 16.586620330810547,
|
| 289 |
+
"logits/rejected": 16.253028869628906,
|
| 290 |
+
"logps/chosen": -0.33331993222236633,
|
| 291 |
+
"logps/rejected": -0.35580095648765564,
|
| 292 |
+
"loss": 0.9827,
|
| 293 |
+
"rewards/accuracies": 0.42500001192092896,
|
| 294 |
+
"rewards/chosen": -0.49997982382774353,
|
| 295 |
+
"rewards/margins": 0.03372158855199814,
|
| 296 |
+
"rewards/rejected": -0.5337014198303223,
|
| 297 |
+
"step": 160
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 0.2668236217382774,
|
| 301 |
+
"grad_norm": 0.09462621062994003,
|
| 302 |
+
"learning_rate": 4.84320497372973e-06,
|
| 303 |
+
"logits/chosen": 16.220312118530273,
|
| 304 |
+
"logits/rejected": 16.105356216430664,
|
| 305 |
+
"logps/chosen": -0.2957116663455963,
|
| 306 |
+
"logps/rejected": -0.3528788089752197,
|
| 307 |
+
"loss": 0.9716,
|
| 308 |
+
"rewards/accuracies": 0.4625000059604645,
|
| 309 |
+
"rewards/chosen": -0.44356757402420044,
|
| 310 |
+
"rewards/margins": 0.08575066924095154,
|
| 311 |
+
"rewards/rejected": -0.5293182134628296,
|
| 312 |
+
"step": 170
|
| 313 |
+
},
|
| 314 |
+
{
|
| 315 |
+
"epoch": 0.28251912889935255,
|
| 316 |
+
"grad_norm": 0.10995513945817947,
|
| 317 |
+
"learning_rate": 4.824441214720629e-06,
|
| 318 |
+
"logits/chosen": 16.811044692993164,
|
| 319 |
+
"logits/rejected": 16.43205451965332,
|
| 320 |
+
"logps/chosen": -0.3263992369174957,
|
| 321 |
+
"logps/rejected": -0.35878634452819824,
|
| 322 |
+
"loss": 0.9546,
|
| 323 |
+
"rewards/accuracies": 0.4124999940395355,
|
| 324 |
+
"rewards/chosen": -0.4895988404750824,
|
| 325 |
+
"rewards/margins": 0.04858064278960228,
|
| 326 |
+
"rewards/rejected": -0.5381795167922974,
|
| 327 |
+
"step": 180
|
| 328 |
+
},
|
| 329 |
+
{
|
| 330 |
+
"epoch": 0.2982146360604277,
|
| 331 |
+
"grad_norm": 0.1348036825656891,
|
| 332 |
+
"learning_rate": 4.804657878971252e-06,
|
| 333 |
+
"logits/chosen": 16.816158294677734,
|
| 334 |
+
"logits/rejected": 16.55413055419922,
|
| 335 |
+
"logps/chosen": -0.31842875480651855,
|
| 336 |
+
"logps/rejected": -0.3791810870170593,
|
| 337 |
+
"loss": 0.9472,
|
| 338 |
+
"rewards/accuracies": 0.5249999761581421,
|
| 339 |
+
"rewards/chosen": -0.47764310240745544,
|
| 340 |
+
"rewards/margins": 0.09112847596406937,
|
| 341 |
+
"rewards/rejected": -0.5687715411186218,
|
| 342 |
+
"step": 190
|
| 343 |
+
},
|
| 344 |
+
{
|
| 345 |
+
"epoch": 0.31391014322150285,
|
| 346 |
+
"grad_norm": 0.13393694162368774,
|
| 347 |
+
"learning_rate": 4.783863644106502e-06,
|
| 348 |
+
"logits/chosen": 17.075885772705078,
|
| 349 |
+
"logits/rejected": 16.982004165649414,
|
| 350 |
+
"logps/chosen": -0.3368823528289795,
|
| 351 |
+
"logps/rejected": -0.3636801540851593,
|
| 352 |
+
"loss": 0.9455,
|
| 353 |
+
"rewards/accuracies": 0.375,
|
| 354 |
+
"rewards/chosen": -0.5053235292434692,
|
| 355 |
+
"rewards/margins": 0.04019671678543091,
|
| 356 |
+
"rewards/rejected": -0.5455202460289001,
|
| 357 |
+
"step": 200
|
| 358 |
+
},
|
| 359 |
+
{
|
| 360 |
+
"epoch": 0.31391014322150285,
|
| 361 |
+
"eval_logits/chosen": 17.17239761352539,
|
| 362 |
+
"eval_logits/rejected": 16.609193801879883,
|
| 363 |
+
"eval_logps/chosen": -0.33508703112602234,
|
| 364 |
+
"eval_logps/rejected": -0.4493505656719208,
|
| 365 |
+
"eval_loss": 0.9283667802810669,
|
| 366 |
+
"eval_rewards/accuracies": 0.4615384638309479,
|
| 367 |
+
"eval_rewards/chosen": -0.5026305913925171,
|
| 368 |
+
"eval_rewards/margins": 0.17139528691768646,
|
| 369 |
+
"eval_rewards/rejected": -0.67402583360672,
|
| 370 |
+
"eval_runtime": 14.5913,
|
| 371 |
+
"eval_samples_per_second": 28.236,
|
| 372 |
+
"eval_steps_per_second": 3.564,
|
| 373 |
+
"step": 200
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.329605650382578,
|
| 377 |
+
"grad_norm": 0.14985737204551697,
|
| 378 |
+
"learning_rate": 4.762067631165049e-06,
|
| 379 |
+
"logits/chosen": 16.92913246154785,
|
| 380 |
+
"logits/rejected": 16.648632049560547,
|
| 381 |
+
"logps/chosen": -0.3326043486595154,
|
| 382 |
+
"logps/rejected": -0.4218205511569977,
|
| 383 |
+
"loss": 0.9225,
|
| 384 |
+
"rewards/accuracies": 0.48750001192092896,
|
| 385 |
+
"rewards/chosen": -0.4989064335823059,
|
| 386 |
+
"rewards/margins": 0.13382436335086823,
|
| 387 |
+
"rewards/rejected": -0.6327308416366577,
|
| 388 |
+
"step": 210
|
| 389 |
+
},
|
| 390 |
+
{
|
| 391 |
+
"epoch": 0.34530115754365315,
|
| 392 |
+
"grad_norm": 0.16972233355045319,
|
| 393 |
+
"learning_rate": 4.7392794005985324e-06,
|
| 394 |
+
"logits/chosen": 17.074562072753906,
|
| 395 |
+
"logits/rejected": 16.863161087036133,
|
| 396 |
+
"logps/chosen": -0.3645358383655548,
|
| 397 |
+
"logps/rejected": -0.5221719145774841,
|
| 398 |
+
"loss": 0.9264,
|
| 399 |
+
"rewards/accuracies": 0.512499988079071,
|
| 400 |
+
"rewards/chosen": -0.5468038320541382,
|
| 401 |
+
"rewards/margins": 0.23645417392253876,
|
| 402 |
+
"rewards/rejected": -0.7832580208778381,
|
| 403 |
+
"step": 220
|
| 404 |
+
},
|
| 405 |
+
{
|
| 406 |
+
"epoch": 0.3609966647047283,
|
| 407 |
+
"grad_norm": 0.20796354115009308,
|
| 408 |
+
"learning_rate": 4.715508948078037e-06,
|
| 409 |
+
"logits/chosen": 17.4168643951416,
|
| 410 |
+
"logits/rejected": 17.00813102722168,
|
| 411 |
+
"logps/chosen": -0.35443297028541565,
|
| 412 |
+
"logps/rejected": -0.5115998983383179,
|
| 413 |
+
"loss": 0.9037,
|
| 414 |
+
"rewards/accuracies": 0.512499988079071,
|
| 415 |
+
"rewards/chosen": -0.5316494703292847,
|
| 416 |
+
"rewards/margins": 0.23575039207935333,
|
| 417 |
+
"rewards/rejected": -0.767399787902832,
|
| 418 |
+
"step": 230
|
| 419 |
+
},
|
| 420 |
+
{
|
| 421 |
+
"epoch": 0.3766921718658034,
|
| 422 |
+
"grad_norm": 0.1932348757982254,
|
| 423 |
+
"learning_rate": 4.690766700109659e-06,
|
| 424 |
+
"logits/chosen": 17.034719467163086,
|
| 425 |
+
"logits/rejected": 16.65166664123535,
|
| 426 |
+
"logps/chosen": -0.3426091969013214,
|
| 427 |
+
"logps/rejected": -0.47884297370910645,
|
| 428 |
+
"loss": 0.9154,
|
| 429 |
+
"rewards/accuracies": 0.5625,
|
| 430 |
+
"rewards/chosen": -0.5139138102531433,
|
| 431 |
+
"rewards/margins": 0.20435063540935516,
|
| 432 |
+
"rewards/rejected": -0.7182644605636597,
|
| 433 |
+
"step": 240
|
| 434 |
+
},
|
| 435 |
+
{
|
| 436 |
+
"epoch": 0.39238767902687854,
|
| 437 |
+
"grad_norm": 1.016450047492981,
|
| 438 |
+
"learning_rate": 4.665063509461098e-06,
|
| 439 |
+
"logits/chosen": 16.935081481933594,
|
| 440 |
+
"logits/rejected": 16.65024185180664,
|
| 441 |
+
"logps/chosen": -0.41738080978393555,
|
| 442 |
+
"logps/rejected": -0.48515433073043823,
|
| 443 |
+
"loss": 0.8774,
|
| 444 |
+
"rewards/accuracies": 0.42500001192092896,
|
| 445 |
+
"rewards/chosen": -0.6260712146759033,
|
| 446 |
+
"rewards/margins": 0.10166029632091522,
|
| 447 |
+
"rewards/rejected": -0.7277315258979797,
|
| 448 |
+
"step": 250
|
| 449 |
+
},
|
| 450 |
+
{
|
| 451 |
+
"epoch": 0.39238767902687854,
|
| 452 |
+
"eval_logits/chosen": 17.221881866455078,
|
| 453 |
+
"eval_logits/rejected": 16.536666870117188,
|
| 454 |
+
"eval_logps/chosen": -0.3716265559196472,
|
| 455 |
+
"eval_logps/rejected": -0.87992262840271,
|
| 456 |
+
"eval_loss": 0.8047741651535034,
|
| 457 |
+
"eval_rewards/accuracies": 0.5,
|
| 458 |
+
"eval_rewards/chosen": -0.5574398636817932,
|
| 459 |
+
"eval_rewards/margins": 0.7624441981315613,
|
| 460 |
+
"eval_rewards/rejected": -1.3198840618133545,
|
| 461 |
+
"eval_runtime": 14.6008,
|
| 462 |
+
"eval_samples_per_second": 28.218,
|
| 463 |
+
"eval_steps_per_second": 3.561,
|
| 464 |
+
"step": 250
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.4080831861879537,
|
| 468 |
+
"grad_norm": 0.2831684648990631,
|
| 469 |
+
"learning_rate": 4.638410650401267e-06,
|
| 470 |
+
"logits/chosen": 17.084096908569336,
|
| 471 |
+
"logits/rejected": 16.564823150634766,
|
| 472 |
+
"logps/chosen": -0.38100525736808777,
|
| 473 |
+
"logps/rejected": -0.8173269033432007,
|
| 474 |
+
"loss": 0.8197,
|
| 475 |
+
"rewards/accuracies": 0.4375,
|
| 476 |
+
"rewards/chosen": -0.5715079307556152,
|
| 477 |
+
"rewards/margins": 0.6544824838638306,
|
| 478 |
+
"rewards/rejected": -1.2259904146194458,
|
| 479 |
+
"step": 260
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 0.42377869334902885,
|
| 483 |
+
"grad_norm": 0.9195305109024048,
|
| 484 |
+
"learning_rate": 4.610819813755038e-06,
|
| 485 |
+
"logits/chosen": 17.316537857055664,
|
| 486 |
+
"logits/rejected": 16.623016357421875,
|
| 487 |
+
"logps/chosen": -0.4713365435600281,
|
| 488 |
+
"logps/rejected": -1.2184536457061768,
|
| 489 |
+
"loss": 0.7493,
|
| 490 |
+
"rewards/accuracies": 0.4375,
|
| 491 |
+
"rewards/chosen": -0.7070047855377197,
|
| 492 |
+
"rewards/margins": 1.120675802230835,
|
| 493 |
+
"rewards/rejected": -1.8276805877685547,
|
| 494 |
+
"step": 270
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 0.439474200510104,
|
| 498 |
+
"grad_norm": 0.5542411804199219,
|
| 499 |
+
"learning_rate": 4.582303101775249e-06,
|
| 500 |
+
"logits/chosen": 17.173168182373047,
|
| 501 |
+
"logits/rejected": 16.759702682495117,
|
| 502 |
+
"logps/chosen": -0.4629506468772888,
|
| 503 |
+
"logps/rejected": -1.3020483255386353,
|
| 504 |
+
"loss": 0.7638,
|
| 505 |
+
"rewards/accuracies": 0.637499988079071,
|
| 506 |
+
"rewards/chosen": -0.6944260597229004,
|
| 507 |
+
"rewards/margins": 1.2586463689804077,
|
| 508 |
+
"rewards/rejected": -1.9530725479125977,
|
| 509 |
+
"step": 280
|
| 510 |
+
},
|
| 511 |
+
{
|
| 512 |
+
"epoch": 0.45516970767117915,
|
| 513 |
+
"grad_norm": 0.2836654782295227,
|
| 514 |
+
"learning_rate": 4.55287302283426e-06,
|
| 515 |
+
"logits/chosen": 17.022424697875977,
|
| 516 |
+
"logits/rejected": 16.43834114074707,
|
| 517 |
+
"logps/chosen": -0.527222216129303,
|
| 518 |
+
"logps/rejected": -1.632965087890625,
|
| 519 |
+
"loss": 0.7466,
|
| 520 |
+
"rewards/accuracies": 0.6625000238418579,
|
| 521 |
+
"rewards/chosen": -0.7908332943916321,
|
| 522 |
+
"rewards/margins": 1.6586145162582397,
|
| 523 |
+
"rewards/rejected": -2.4494476318359375,
|
| 524 |
+
"step": 290
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 0.47086521483225424,
|
| 528 |
+
"grad_norm": 0.658358097076416,
|
| 529 |
+
"learning_rate": 4.522542485937369e-06,
|
| 530 |
+
"logits/chosen": 16.6541690826416,
|
| 531 |
+
"logits/rejected": 16.34103012084961,
|
| 532 |
+
"logps/chosen": -0.50589519739151,
|
| 533 |
+
"logps/rejected": -1.8910831212997437,
|
| 534 |
+
"loss": 0.6982,
|
| 535 |
+
"rewards/accuracies": 0.699999988079071,
|
| 536 |
+
"rewards/chosen": -0.7588427662849426,
|
| 537 |
+
"rewards/margins": 2.077782154083252,
|
| 538 |
+
"rewards/rejected": -2.8366246223449707,
|
| 539 |
+
"step": 300
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"epoch": 0.47086521483225424,
|
| 543 |
+
"eval_logits/chosen": 16.950870513916016,
|
| 544 |
+
"eval_logits/rejected": 16.154659271240234,
|
| 545 |
+
"eval_logps/chosen": -0.5594518184661865,
|
| 546 |
+
"eval_logps/rejected": -1.9617934226989746,
|
| 547 |
+
"eval_loss": 0.7422243356704712,
|
| 548 |
+
"eval_rewards/accuracies": 0.7115384340286255,
|
| 549 |
+
"eval_rewards/chosen": -0.839177668094635,
|
| 550 |
+
"eval_rewards/margins": 2.103512763977051,
|
| 551 |
+
"eval_rewards/rejected": -2.942690372467041,
|
| 552 |
+
"eval_runtime": 14.5969,
|
| 553 |
+
"eval_samples_per_second": 28.225,
|
| 554 |
+
"eval_steps_per_second": 3.562,
|
| 555 |
+
"step": 300
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.4865607219933294,
|
| 559 |
+
"grad_norm": 1.1044840812683105,
|
| 560 |
+
"learning_rate": 4.491324795060491e-06,
|
| 561 |
+
"logits/chosen": 17.233802795410156,
|
| 562 |
+
"logits/rejected": 16.292484283447266,
|
| 563 |
+
"logps/chosen": -0.5940151214599609,
|
| 564 |
+
"logps/rejected": -2.041588306427002,
|
| 565 |
+
"loss": 0.7415,
|
| 566 |
+
"rewards/accuracies": 0.6625000238418579,
|
| 567 |
+
"rewards/chosen": -0.8910226821899414,
|
| 568 |
+
"rewards/margins": 2.1713600158691406,
|
| 569 |
+
"rewards/rejected": -3.062382459640503,
|
| 570 |
+
"step": 310
|
| 571 |
+
},
|
| 572 |
+
{
|
| 573 |
+
"epoch": 0.5022562291544046,
|
| 574 |
+
"grad_norm": 0.524356484413147,
|
| 575 |
+
"learning_rate": 4.4592336433146e-06,
|
| 576 |
+
"logits/chosen": 17.155719757080078,
|
| 577 |
+
"logits/rejected": 16.328996658325195,
|
| 578 |
+
"logps/chosen": -0.6095865964889526,
|
| 579 |
+
"logps/rejected": -1.831730604171753,
|
| 580 |
+
"loss": 0.7249,
|
| 581 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 582 |
+
"rewards/chosen": -0.9143797755241394,
|
| 583 |
+
"rewards/margins": 1.8332160711288452,
|
| 584 |
+
"rewards/rejected": -2.747596025466919,
|
| 585 |
+
"step": 320
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 0.5179517363154797,
|
| 589 |
+
"grad_norm": 0.4383145570755005,
|
| 590 |
+
"learning_rate": 4.426283106939474e-06,
|
| 591 |
+
"logits/chosen": 16.774551391601562,
|
| 592 |
+
"logits/rejected": 16.532238006591797,
|
| 593 |
+
"logps/chosen": -0.669482946395874,
|
| 594 |
+
"logps/rejected": -1.7058660984039307,
|
| 595 |
+
"loss": 0.6868,
|
| 596 |
+
"rewards/accuracies": 0.6625000238418579,
|
| 597 |
+
"rewards/chosen": -1.004224419593811,
|
| 598 |
+
"rewards/margins": 1.5545748472213745,
|
| 599 |
+
"rewards/rejected": -2.5587992668151855,
|
| 600 |
+
"step": 330
|
| 601 |
+
},
|
| 602 |
+
{
|
| 603 |
+
"epoch": 0.5336472434765548,
|
| 604 |
+
"grad_norm": 3.194026231765747,
|
| 605 |
+
"learning_rate": 4.3924876391293915e-06,
|
| 606 |
+
"logits/chosen": 16.963787078857422,
|
| 607 |
+
"logits/rejected": 16.3092041015625,
|
| 608 |
+
"logps/chosen": -0.7091449499130249,
|
| 609 |
+
"logps/rejected": -2.1812453269958496,
|
| 610 |
+
"loss": 0.706,
|
| 611 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 612 |
+
"rewards/chosen": -1.0637174844741821,
|
| 613 |
+
"rewards/margins": 2.208150863647461,
|
| 614 |
+
"rewards/rejected": -3.2718684673309326,
|
| 615 |
+
"step": 340
|
| 616 |
+
},
|
| 617 |
+
{
|
| 618 |
+
"epoch": 0.5493427506376299,
|
| 619 |
+
"grad_norm": 0.5639687776565552,
|
| 620 |
+
"learning_rate": 4.357862063693486e-06,
|
| 621 |
+
"logits/chosen": 17.019062042236328,
|
| 622 |
+
"logits/rejected": 16.315940856933594,
|
| 623 |
+
"logps/chosen": -0.8996315002441406,
|
| 624 |
+
"logps/rejected": -2.0307648181915283,
|
| 625 |
+
"loss": 0.674,
|
| 626 |
+
"rewards/accuracies": 0.75,
|
| 627 |
+
"rewards/chosen": -1.349447250366211,
|
| 628 |
+
"rewards/margins": 1.696699857711792,
|
| 629 |
+
"rewards/rejected": -3.046147108078003,
|
| 630 |
+
"step": 350
|
| 631 |
+
},
|
| 632 |
+
{
|
| 633 |
+
"epoch": 0.5493427506376299,
|
| 634 |
+
"eval_logits/chosen": 16.70261001586914,
|
| 635 |
+
"eval_logits/rejected": 15.845681190490723,
|
| 636 |
+
"eval_logps/chosen": -0.9138904809951782,
|
| 637 |
+
"eval_logps/rejected": -2.5003392696380615,
|
| 638 |
+
"eval_loss": 0.6938430070877075,
|
| 639 |
+
"eval_rewards/accuracies": 0.8461538553237915,
|
| 640 |
+
"eval_rewards/chosen": -1.3708356618881226,
|
| 641 |
+
"eval_rewards/margins": 2.3796732425689697,
|
| 642 |
+
"eval_rewards/rejected": -3.75050950050354,
|
| 643 |
+
"eval_runtime": 14.6072,
|
| 644 |
+
"eval_samples_per_second": 28.205,
|
| 645 |
+
"eval_steps_per_second": 3.56,
|
| 646 |
+
"step": 350
|
| 647 |
+
},
|
| 648 |
+
{
|
| 649 |
+
"epoch": 0.5650382577987051,
|
| 650 |
+
"grad_norm": 0.6917555928230286,
|
| 651 |
+
"learning_rate": 4.322421568553529e-06,
|
| 652 |
+
"logits/chosen": 16.269399642944336,
|
| 653 |
+
"logits/rejected": 15.73193645477295,
|
| 654 |
+
"logps/chosen": -1.0171478986740112,
|
| 655 |
+
"logps/rejected": -2.201298236846924,
|
| 656 |
+
"loss": 0.651,
|
| 657 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 658 |
+
"rewards/chosen": -1.525721788406372,
|
| 659 |
+
"rewards/margins": 1.7762253284454346,
|
| 660 |
+
"rewards/rejected": -3.3019473552703857,
|
| 661 |
+
"step": 360
|
| 662 |
+
},
|
| 663 |
+
{
|
| 664 |
+
"epoch": 0.5807337649597802,
|
| 665 |
+
"grad_norm": 1.151584267616272,
|
| 666 |
+
"learning_rate": 4.286181699082008e-06,
|
| 667 |
+
"logits/chosen": 15.88188648223877,
|
| 668 |
+
"logits/rejected": 15.261810302734375,
|
| 669 |
+
"logps/chosen": -1.2272206544876099,
|
| 670 |
+
"logps/rejected": -3.0203287601470947,
|
| 671 |
+
"loss": 0.665,
|
| 672 |
+
"rewards/accuracies": 0.800000011920929,
|
| 673 |
+
"rewards/chosen": -1.8408310413360596,
|
| 674 |
+
"rewards/margins": 2.689661741256714,
|
| 675 |
+
"rewards/rejected": -4.530492305755615,
|
| 676 |
+
"step": 370
|
| 677 |
+
},
|
| 678 |
+
{
|
| 679 |
+
"epoch": 0.5964292721208554,
|
| 680 |
+
"grad_norm": 1.1879621744155884,
|
| 681 |
+
"learning_rate": 4.249158351283414e-06,
|
| 682 |
+
"logits/chosen": 15.906834602355957,
|
| 683 |
+
"logits/rejected": 15.42895221710205,
|
| 684 |
+
"logps/chosen": -1.5027008056640625,
|
| 685 |
+
"logps/rejected": -2.9172558784484863,
|
| 686 |
+
"loss": 0.5998,
|
| 687 |
+
"rewards/accuracies": 0.887499988079071,
|
| 688 |
+
"rewards/chosen": -2.2540509700775146,
|
| 689 |
+
"rewards/margins": 2.1218326091766357,
|
| 690 |
+
"rewards/rejected": -4.37588357925415,
|
| 691 |
+
"step": 380
|
| 692 |
+
},
|
| 693 |
+
{
|
| 694 |
+
"epoch": 0.6121247792819305,
|
| 695 |
+
"grad_norm": 1.678884744644165,
|
| 696 |
+
"learning_rate": 4.211367764821722e-06,
|
| 697 |
+
"logits/chosen": 15.630645751953125,
|
| 698 |
+
"logits/rejected": 14.888415336608887,
|
| 699 |
+
"logps/chosen": -2.231367349624634,
|
| 700 |
+
"logps/rejected": -3.5754780769348145,
|
| 701 |
+
"loss": 0.5937,
|
| 702 |
+
"rewards/accuracies": 0.824999988079071,
|
| 703 |
+
"rewards/chosen": -3.347050905227661,
|
| 704 |
+
"rewards/margins": 2.0161664485931396,
|
| 705 |
+
"rewards/rejected": -5.363216876983643,
|
| 706 |
+
"step": 390
|
| 707 |
+
},
|
| 708 |
+
{
|
| 709 |
+
"epoch": 0.6278202864430057,
|
| 710 |
+
"grad_norm": 2.008671998977661,
|
| 711 |
+
"learning_rate": 4.172826515897146e-06,
|
| 712 |
+
"logits/chosen": 15.024614334106445,
|
| 713 |
+
"logits/rejected": 14.981277465820312,
|
| 714 |
+
"logps/chosen": -2.665980815887451,
|
| 715 |
+
"logps/rejected": -3.582763195037842,
|
| 716 |
+
"loss": 0.5722,
|
| 717 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 718 |
+
"rewards/chosen": -3.998971462249756,
|
| 719 |
+
"rewards/margins": 1.3751739263534546,
|
| 720 |
+
"rewards/rejected": -5.374145030975342,
|
| 721 |
+
"step": 400
|
| 722 |
+
},
|
| 723 |
+
{
|
| 724 |
+
"epoch": 0.6278202864430057,
|
| 725 |
+
"eval_logits/chosen": 15.45953369140625,
|
| 726 |
+
"eval_logits/rejected": 14.663763046264648,
|
| 727 |
+
"eval_logps/chosen": -2.7269070148468018,
|
| 728 |
+
"eval_logps/rejected": -4.235719680786133,
|
| 729 |
+
"eval_loss": 0.5668805837631226,
|
| 730 |
+
"eval_rewards/accuracies": 0.8846153616905212,
|
| 731 |
+
"eval_rewards/chosen": -4.090360641479492,
|
| 732 |
+
"eval_rewards/margins": 2.263219118118286,
|
| 733 |
+
"eval_rewards/rejected": -6.353579998016357,
|
| 734 |
+
"eval_runtime": 14.6036,
|
| 735 |
+
"eval_samples_per_second": 28.212,
|
| 736 |
+
"eval_steps_per_second": 3.561,
|
| 737 |
+
"step": 400
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 0.6435157936040808,
|
| 741 |
+
"grad_norm": 1.018362045288086,
|
| 742 |
+
"learning_rate": 4.133551509975264e-06,
|
| 743 |
+
"logits/chosen": 15.027656555175781,
|
| 744 |
+
"logits/rejected": 14.74786376953125,
|
| 745 |
+
"logps/chosen": -2.76598858833313,
|
| 746 |
+
"logps/rejected": -4.335003852844238,
|
| 747 |
+
"loss": 0.5412,
|
| 748 |
+
"rewards/accuracies": 0.8374999761581421,
|
| 749 |
+
"rewards/chosen": -4.148982524871826,
|
| 750 |
+
"rewards/margins": 2.3535237312316895,
|
| 751 |
+
"rewards/rejected": -6.502506256103516,
|
| 752 |
+
"step": 410
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 0.659211300765156,
|
| 756 |
+
"grad_norm": 2.2899765968322754,
|
| 757 |
+
"learning_rate": 4.093559974371725e-06,
|
| 758 |
+
"logits/chosen": 15.15197467803955,
|
| 759 |
+
"logits/rejected": 14.681947708129883,
|
| 760 |
+
"logps/chosen": -2.9831995964050293,
|
| 761 |
+
"logps/rejected": -4.723050117492676,
|
| 762 |
+
"loss": 0.4652,
|
| 763 |
+
"rewards/accuracies": 0.862500011920929,
|
| 764 |
+
"rewards/chosen": -4.474799633026123,
|
| 765 |
+
"rewards/margins": 2.6097757816314697,
|
| 766 |
+
"rewards/rejected": -7.0845746994018555,
|
| 767 |
+
"step": 420
|
| 768 |
+
},
|
| 769 |
+
{
|
| 770 |
+
"epoch": 0.6749068079262311,
|
| 771 |
+
"grad_norm": 1.419196605682373,
|
| 772 |
+
"learning_rate": 4.052869450695776e-06,
|
| 773 |
+
"logits/chosen": 15.000317573547363,
|
| 774 |
+
"logits/rejected": 14.538591384887695,
|
| 775 |
+
"logps/chosen": -3.543004274368286,
|
| 776 |
+
"logps/rejected": -5.037484169006348,
|
| 777 |
+
"loss": 0.427,
|
| 778 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 779 |
+
"rewards/chosen": -5.314507007598877,
|
| 780 |
+
"rewards/margins": 2.2417192459106445,
|
| 781 |
+
"rewards/rejected": -7.556225776672363,
|
| 782 |
+
"step": 430
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 0.6906023150873063,
|
| 786 |
+
"grad_norm": 1.639215111732483,
|
| 787 |
+
"learning_rate": 4.011497787155938e-06,
|
| 788 |
+
"logits/chosen": 14.798017501831055,
|
| 789 |
+
"logits/rejected": 14.236108779907227,
|
| 790 |
+
"logps/chosen": -4.154418468475342,
|
| 791 |
+
"logps/rejected": -6.0606184005737305,
|
| 792 |
+
"loss": 0.4692,
|
| 793 |
+
"rewards/accuracies": 0.8374999761581421,
|
| 794 |
+
"rewards/chosen": -6.231626987457275,
|
| 795 |
+
"rewards/margins": 2.859299421310425,
|
| 796 |
+
"rewards/rejected": -9.090926170349121,
|
| 797 |
+
"step": 440
|
| 798 |
+
},
|
| 799 |
+
{
|
| 800 |
+
"epoch": 0.7062978222483814,
|
| 801 |
+
"grad_norm": 1.6229957342147827,
|
| 802 |
+
"learning_rate": 3.969463130731183e-06,
|
| 803 |
+
"logits/chosen": 14.495908737182617,
|
| 804 |
+
"logits/rejected": 13.978658676147461,
|
| 805 |
+
"logps/chosen": -3.4806721210479736,
|
| 806 |
+
"logps/rejected": -5.746390342712402,
|
| 807 |
+
"loss": 0.4218,
|
| 808 |
+
"rewards/accuracies": 0.8500000238418579,
|
| 809 |
+
"rewards/chosen": -5.221007823944092,
|
| 810 |
+
"rewards/margins": 3.3985772132873535,
|
| 811 |
+
"rewards/rejected": -8.619585037231445,
|
| 812 |
+
"step": 450
|
| 813 |
+
},
|
| 814 |
+
{
|
| 815 |
+
"epoch": 0.7062978222483814,
|
| 816 |
+
"eval_logits/chosen": 15.033548355102539,
|
| 817 |
+
"eval_logits/rejected": 14.20669174194336,
|
| 818 |
+
"eval_logps/chosen": -3.423999547958374,
|
| 819 |
+
"eval_logps/rejected": -5.855647563934326,
|
| 820 |
+
"eval_loss": 0.45794492959976196,
|
| 821 |
+
"eval_rewards/accuracies": 0.942307710647583,
|
| 822 |
+
"eval_rewards/chosen": -5.135998725891113,
|
| 823 |
+
"eval_rewards/margins": 3.6474733352661133,
|
| 824 |
+
"eval_rewards/rejected": -8.783472061157227,
|
| 825 |
+
"eval_runtime": 14.6088,
|
| 826 |
+
"eval_samples_per_second": 28.202,
|
| 827 |
+
"eval_steps_per_second": 3.559,
|
| 828 |
+
"step": 450
|
| 829 |
+
},
|
| 830 |
+
{
|
| 831 |
+
"epoch": 0.7219933294094566,
|
| 832 |
+
"grad_norm": 1.7993106842041016,
|
| 833 |
+
"learning_rate": 3.92678391921108e-06,
|
| 834 |
+
"logits/chosen": 14.569323539733887,
|
| 835 |
+
"logits/rejected": 13.937005996704102,
|
| 836 |
+
"logps/chosen": -3.3623528480529785,
|
| 837 |
+
"logps/rejected": -5.784353733062744,
|
| 838 |
+
"loss": 0.3675,
|
| 839 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 840 |
+
"rewards/chosen": -5.043529033660889,
|
| 841 |
+
"rewards/margins": 3.6330013275146484,
|
| 842 |
+
"rewards/rejected": -8.676530838012695,
|
| 843 |
+
"step": 460
|
| 844 |
+
},
|
| 845 |
+
{
|
| 846 |
+
"epoch": 0.7376888365705316,
|
| 847 |
+
"grad_norm": 2.1828088760375977,
|
| 848 |
+
"learning_rate": 3.88347887310836e-06,
|
| 849 |
+
"logits/chosen": 13.90056037902832,
|
| 850 |
+
"logits/rejected": 13.489587783813477,
|
| 851 |
+
"logps/chosen": -3.959341049194336,
|
| 852 |
+
"logps/rejected": -6.217314720153809,
|
| 853 |
+
"loss": 0.3653,
|
| 854 |
+
"rewards/accuracies": 0.862500011920929,
|
| 855 |
+
"rewards/chosen": -5.939011573791504,
|
| 856 |
+
"rewards/margins": 3.3869614601135254,
|
| 857 |
+
"rewards/rejected": -9.325971603393555,
|
| 858 |
+
"step": 470
|
| 859 |
+
},
|
| 860 |
+
{
|
| 861 |
+
"epoch": 0.7533843437316068,
|
| 862 |
+
"grad_norm": 2.128547191619873,
|
| 863 |
+
"learning_rate": 3.839566987447492e-06,
|
| 864 |
+
"logits/chosen": 14.113011360168457,
|
| 865 |
+
"logits/rejected": 13.571691513061523,
|
| 866 |
+
"logps/chosen": -4.004420280456543,
|
| 867 |
+
"logps/rejected": -6.216452121734619,
|
| 868 |
+
"loss": 0.3439,
|
| 869 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 870 |
+
"rewards/chosen": -6.0066304206848145,
|
| 871 |
+
"rewards/margins": 3.3180477619171143,
|
| 872 |
+
"rewards/rejected": -9.324677467346191,
|
| 873 |
+
"step": 480
|
| 874 |
+
},
|
| 875 |
+
{
|
| 876 |
+
"epoch": 0.7690798508926819,
|
| 877 |
+
"grad_norm": 3.2651569843292236,
|
| 878 |
+
"learning_rate": 3.795067523432826e-06,
|
| 879 |
+
"logits/chosen": 13.918218612670898,
|
| 880 |
+
"logits/rejected": 13.147412300109863,
|
| 881 |
+
"logps/chosen": -4.574521064758301,
|
| 882 |
+
"logps/rejected": -7.411087989807129,
|
| 883 |
+
"loss": 0.3645,
|
| 884 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 885 |
+
"rewards/chosen": -6.861782073974609,
|
| 886 |
+
"rewards/margins": 4.254849910736084,
|
| 887 |
+
"rewards/rejected": -11.116632461547852,
|
| 888 |
+
"step": 490
|
| 889 |
+
},
|
| 890 |
+
{
|
| 891 |
+
"epoch": 0.7847753580537571,
|
| 892 |
+
"grad_norm": 2.163595676422119,
|
| 893 |
+
"learning_rate": 3.7500000000000005e-06,
|
| 894 |
+
"logits/chosen": 13.972379684448242,
|
| 895 |
+
"logits/rejected": 13.390825271606445,
|
| 896 |
+
"logps/chosen": -3.931866407394409,
|
| 897 |
+
"logps/rejected": -6.547454833984375,
|
| 898 |
+
"loss": 0.3698,
|
| 899 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 900 |
+
"rewards/chosen": -5.897799491882324,
|
| 901 |
+
"rewards/margins": 3.923382520675659,
|
| 902 |
+
"rewards/rejected": -9.821184158325195,
|
| 903 |
+
"step": 500
|
| 904 |
+
},
|
| 905 |
+
{
|
| 906 |
+
"epoch": 0.7847753580537571,
|
| 907 |
+
"eval_logits/chosen": 14.325434684753418,
|
| 908 |
+
"eval_logits/rejected": 13.4517822265625,
|
| 909 |
+
"eval_logps/chosen": -3.7717440128326416,
|
| 910 |
+
"eval_logps/rejected": -6.710938453674316,
|
| 911 |
+
"eval_loss": 0.38915327191352844,
|
| 912 |
+
"eval_rewards/accuracies": 0.942307710647583,
|
| 913 |
+
"eval_rewards/chosen": -5.657615661621094,
|
| 914 |
+
"eval_rewards/margins": 4.408792018890381,
|
| 915 |
+
"eval_rewards/rejected": -10.066408157348633,
|
| 916 |
+
"eval_runtime": 14.5996,
|
| 917 |
+
"eval_samples_per_second": 28.22,
|
| 918 |
+
"eval_steps_per_second": 3.562,
|
| 919 |
+
"step": 500
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"epoch": 0.8004708652148322,
|
| 923 |
+
"grad_norm": 2.7098453044891357,
|
| 924 |
+
"learning_rate": 3.7043841852542884e-06,
|
| 925 |
+
"logits/chosen": 13.909749984741211,
|
| 926 |
+
"logits/rejected": 13.060315132141113,
|
| 927 |
+
"logps/chosen": -3.848449230194092,
|
| 928 |
+
"logps/rejected": -6.747067928314209,
|
| 929 |
+
"loss": 0.3547,
|
| 930 |
+
"rewards/accuracies": 0.875,
|
| 931 |
+
"rewards/chosen": -5.772673606872559,
|
| 932 |
+
"rewards/margins": 4.347928524017334,
|
| 933 |
+
"rewards/rejected": -10.120603561401367,
|
| 934 |
+
"step": 510
|
| 935 |
+
},
|
| 936 |
+
{
|
| 937 |
+
"epoch": 0.8161663723759074,
|
| 938 |
+
"grad_norm": 1.693467140197754,
|
| 939 |
+
"learning_rate": 3.658240087799655e-06,
|
| 940 |
+
"logits/chosen": 13.52336311340332,
|
| 941 |
+
"logits/rejected": 13.026924133300781,
|
| 942 |
+
"logps/chosen": -4.218991279602051,
|
| 943 |
+
"logps/rejected": -6.984036445617676,
|
| 944 |
+
"loss": 0.3718,
|
| 945 |
+
"rewards/accuracies": 0.9375,
|
| 946 |
+
"rewards/chosen": -6.328486442565918,
|
| 947 |
+
"rewards/margins": 4.1475677490234375,
|
| 948 |
+
"rewards/rejected": -10.476054191589355,
|
| 949 |
+
"step": 520
|
| 950 |
+
},
|
| 951 |
+
{
|
| 952 |
+
"epoch": 0.8318618795369825,
|
| 953 |
+
"grad_norm": 2.4496915340423584,
|
| 954 |
+
"learning_rate": 3.611587947962319e-06,
|
| 955 |
+
"logits/chosen": 13.840197563171387,
|
| 956 |
+
"logits/rejected": 13.181543350219727,
|
| 957 |
+
"logps/chosen": -4.708582401275635,
|
| 958 |
+
"logps/rejected": -7.19864559173584,
|
| 959 |
+
"loss": 0.329,
|
| 960 |
+
"rewards/accuracies": 0.925000011920929,
|
| 961 |
+
"rewards/chosen": -7.062872886657715,
|
| 962 |
+
"rewards/margins": 3.7350947856903076,
|
| 963 |
+
"rewards/rejected": -10.797967910766602,
|
| 964 |
+
"step": 530
|
| 965 |
+
},
|
| 966 |
+
{
|
| 967 |
+
"epoch": 0.8475573866980577,
|
| 968 |
+
"grad_norm": 5.772508144378662,
|
| 969 |
+
"learning_rate": 3.564448228912682e-06,
|
| 970 |
+
"logits/chosen": 13.432130813598633,
|
| 971 |
+
"logits/rejected": 12.825490951538086,
|
| 972 |
+
"logps/chosen": -4.506819248199463,
|
| 973 |
+
"logps/rejected": -7.506214141845703,
|
| 974 |
+
"loss": 0.2544,
|
| 975 |
+
"rewards/accuracies": 0.925000011920929,
|
| 976 |
+
"rewards/chosen": -6.760228633880615,
|
| 977 |
+
"rewards/margins": 4.499091625213623,
|
| 978 |
+
"rewards/rejected": -11.259321212768555,
|
| 979 |
+
"step": 540
|
| 980 |
+
},
|
| 981 |
+
{
|
| 982 |
+
"epoch": 0.8632528938591328,
|
| 983 |
+
"grad_norm": 1.7776503562927246,
|
| 984 |
+
"learning_rate": 3.516841607689501e-06,
|
| 985 |
+
"logits/chosen": 13.536295890808105,
|
| 986 |
+
"logits/rejected": 12.703726768493652,
|
| 987 |
+
"logps/chosen": -4.724976062774658,
|
| 988 |
+
"logps/rejected": -7.864454746246338,
|
| 989 |
+
"loss": 0.2795,
|
| 990 |
+
"rewards/accuracies": 0.925000011920929,
|
| 991 |
+
"rewards/chosen": -7.087464332580566,
|
| 992 |
+
"rewards/margins": 4.7092180252075195,
|
| 993 |
+
"rewards/rejected": -11.796682357788086,
|
| 994 |
+
"step": 550
|
| 995 |
+
},
|
| 996 |
+
{
|
| 997 |
+
"epoch": 0.8632528938591328,
|
| 998 |
+
"eval_logits/chosen": 13.692171096801758,
|
| 999 |
+
"eval_logits/rejected": 12.81371021270752,
|
| 1000 |
+
"eval_logps/chosen": -4.32358980178833,
|
| 1001 |
+
"eval_logps/rejected": -7.658167839050293,
|
| 1002 |
+
"eval_loss": 0.335475891828537,
|
| 1003 |
+
"eval_rewards/accuracies": 0.942307710647583,
|
| 1004 |
+
"eval_rewards/chosen": -6.485384464263916,
|
| 1005 |
+
"eval_rewards/margins": 5.00186824798584,
|
| 1006 |
+
"eval_rewards/rejected": -11.487252235412598,
|
| 1007 |
+
"eval_runtime": 14.6093,
|
| 1008 |
+
"eval_samples_per_second": 28.201,
|
| 1009 |
+
"eval_steps_per_second": 3.559,
|
| 1010 |
+
"step": 550
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 0.878948401020208,
|
| 1014 |
+
"grad_norm": 1.7613197565078735,
|
| 1015 |
+
"learning_rate": 3.4687889661302577e-06,
|
| 1016 |
+
"logits/chosen": 13.689091682434082,
|
| 1017 |
+
"logits/rejected": 12.911500930786133,
|
| 1018 |
+
"logps/chosen": -3.8375942707061768,
|
| 1019 |
+
"logps/rejected": -6.98319149017334,
|
| 1020 |
+
"loss": 0.3008,
|
| 1021 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 1022 |
+
"rewards/chosen": -5.7563910484313965,
|
| 1023 |
+
"rewards/margins": 4.7183966636657715,
|
| 1024 |
+
"rewards/rejected": -10.474787712097168,
|
| 1025 |
+
"step": 560
|
| 1026 |
+
},
|
| 1027 |
+
{
|
| 1028 |
+
"epoch": 0.8946439081812831,
|
| 1029 |
+
"grad_norm": 2.1285247802734375,
|
| 1030 |
+
"learning_rate": 3.4203113817116955e-06,
|
| 1031 |
+
"logits/chosen": 13.902356147766113,
|
| 1032 |
+
"logits/rejected": 12.766775131225586,
|
| 1033 |
+
"logps/chosen": -3.7727108001708984,
|
| 1034 |
+
"logps/rejected": -7.6362409591674805,
|
| 1035 |
+
"loss": 0.2816,
|
| 1036 |
+
"rewards/accuracies": 0.9375,
|
| 1037 |
+
"rewards/chosen": -5.6590657234191895,
|
| 1038 |
+
"rewards/margins": 5.795297622680664,
|
| 1039 |
+
"rewards/rejected": -11.454363822937012,
|
| 1040 |
+
"step": 570
|
| 1041 |
+
},
|
| 1042 |
+
{
|
| 1043 |
+
"epoch": 0.9103394153423583,
|
| 1044 |
+
"grad_norm": 1.6965669393539429,
|
| 1045 |
+
"learning_rate": 3.3714301183045382e-06,
|
| 1046 |
+
"logits/chosen": 13.374112129211426,
|
| 1047 |
+
"logits/rejected": 12.414613723754883,
|
| 1048 |
+
"logps/chosen": -4.747832298278809,
|
| 1049 |
+
"logps/rejected": -8.099973678588867,
|
| 1050 |
+
"loss": 0.2921,
|
| 1051 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 1052 |
+
"rewards/chosen": -7.121748924255371,
|
| 1053 |
+
"rewards/margins": 5.028212547302246,
|
| 1054 |
+
"rewards/rejected": -12.1499605178833,
|
| 1055 |
+
"step": 580
|
| 1056 |
+
},
|
| 1057 |
+
{
|
| 1058 |
+
"epoch": 0.9260349225034334,
|
| 1059 |
+
"grad_norm": 2.5470242500305176,
|
| 1060 |
+
"learning_rate": 3.3221666168464584e-06,
|
| 1061 |
+
"logits/chosen": 13.287660598754883,
|
| 1062 |
+
"logits/rejected": 12.554139137268066,
|
| 1063 |
+
"logps/chosen": -4.818625450134277,
|
| 1064 |
+
"logps/rejected": -8.005637168884277,
|
| 1065 |
+
"loss": 0.3155,
|
| 1066 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1067 |
+
"rewards/chosen": -7.227938175201416,
|
| 1068 |
+
"rewards/margins": 4.780518531799316,
|
| 1069 |
+
"rewards/rejected": -12.008456230163574,
|
| 1070 |
+
"step": 590
|
| 1071 |
+
},
|
| 1072 |
+
{
|
| 1073 |
+
"epoch": 0.9417304296645085,
|
| 1074 |
+
"grad_norm": 2.374859571456909,
|
| 1075 |
+
"learning_rate": 3.272542485937369e-06,
|
| 1076 |
+
"logits/chosen": 13.152572631835938,
|
| 1077 |
+
"logits/rejected": 12.557012557983398,
|
| 1078 |
+
"logps/chosen": -3.7581982612609863,
|
| 1079 |
+
"logps/rejected": -6.985205173492432,
|
| 1080 |
+
"loss": 0.2565,
|
| 1081 |
+
"rewards/accuracies": 0.9375,
|
| 1082 |
+
"rewards/chosen": -5.6372971534729,
|
| 1083 |
+
"rewards/margins": 4.840510368347168,
|
| 1084 |
+
"rewards/rejected": -10.477807998657227,
|
| 1085 |
+
"step": 600
|
| 1086 |
+
},
|
| 1087 |
+
{
|
| 1088 |
+
"epoch": 0.9417304296645085,
|
| 1089 |
+
"eval_logits/chosen": 13.585193634033203,
|
| 1090 |
+
"eval_logits/rejected": 12.657953262329102,
|
| 1091 |
+
"eval_logps/chosen": -4.077207088470459,
|
| 1092 |
+
"eval_logps/rejected": -7.660572528839111,
|
| 1093 |
+
"eval_loss": 0.30504748225212097,
|
| 1094 |
+
"eval_rewards/accuracies": 0.9615384340286255,
|
| 1095 |
+
"eval_rewards/chosen": -6.115810871124268,
|
| 1096 |
+
"eval_rewards/margins": 5.375046730041504,
|
| 1097 |
+
"eval_rewards/rejected": -11.49085807800293,
|
| 1098 |
+
"eval_runtime": 14.6046,
|
| 1099 |
+
"eval_samples_per_second": 28.21,
|
| 1100 |
+
"eval_steps_per_second": 3.561,
|
| 1101 |
+
"step": 600
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"epoch": 0.9574259368255836,
|
| 1105 |
+
"grad_norm": 3.3574252128601074,
|
| 1106 |
+
"learning_rate": 3.222579492361179e-06,
|
| 1107 |
+
"logits/chosen": 12.703638076782227,
|
| 1108 |
+
"logits/rejected": 11.910614013671875,
|
| 1109 |
+
"logps/chosen": -4.188838005065918,
|
| 1110 |
+
"logps/rejected": -7.538239479064941,
|
| 1111 |
+
"loss": 0.2527,
|
| 1112 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 1113 |
+
"rewards/chosen": -6.283257961273193,
|
| 1114 |
+
"rewards/margins": 5.024100303649902,
|
| 1115 |
+
"rewards/rejected": -11.30735969543457,
|
| 1116 |
+
"step": 610
|
| 1117 |
+
},
|
| 1118 |
+
{
|
| 1119 |
+
"epoch": 0.9731214439866588,
|
| 1120 |
+
"grad_norm": 1.726338505744934,
|
| 1121 |
+
"learning_rate": 3.1722995515381644e-06,
|
| 1122 |
+
"logits/chosen": 13.002418518066406,
|
| 1123 |
+
"logits/rejected": 12.060869216918945,
|
| 1124 |
+
"logps/chosen": -4.711850166320801,
|
| 1125 |
+
"logps/rejected": -8.62987995147705,
|
| 1126 |
+
"loss": 0.2895,
|
| 1127 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 1128 |
+
"rewards/chosen": -7.067776679992676,
|
| 1129 |
+
"rewards/margins": 5.877043724060059,
|
| 1130 |
+
"rewards/rejected": -12.944819450378418,
|
| 1131 |
+
"step": 620
|
| 1132 |
+
},
|
| 1133 |
+
{
|
| 1134 |
+
"epoch": 0.9888169511477339,
|
| 1135 |
+
"grad_norm": 2.5146796703338623,
|
| 1136 |
+
"learning_rate": 3.121724717912138e-06,
|
| 1137 |
+
"logits/chosen": 13.190832138061523,
|
| 1138 |
+
"logits/rejected": 11.927602767944336,
|
| 1139 |
+
"logps/chosen": -4.228806495666504,
|
| 1140 |
+
"logps/rejected": -8.436616897583008,
|
| 1141 |
+
"loss": 0.2284,
|
| 1142 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 1143 |
+
"rewards/chosen": -6.343210220336914,
|
| 1144 |
+
"rewards/margins": 6.311715602874756,
|
| 1145 |
+
"rewards/rejected": -12.654926300048828,
|
| 1146 |
+
"step": 630
|
| 1147 |
+
},
|
| 1148 |
+
{
|
| 1149 |
+
"epoch": 1.003139101432215,
|
| 1150 |
+
"grad_norm": 4.229348659515381,
|
| 1151 |
+
"learning_rate": 3.0708771752766397e-06,
|
| 1152 |
+
"logits/chosen": 13.159438133239746,
|
| 1153 |
+
"logits/rejected": 12.417133331298828,
|
| 1154 |
+
"logps/chosen": -4.084325313568115,
|
| 1155 |
+
"logps/rejected": -7.705818176269531,
|
| 1156 |
+
"loss": 0.2276,
|
| 1157 |
+
"rewards/accuracies": 0.9178082346916199,
|
| 1158 |
+
"rewards/chosen": -6.126489162445068,
|
| 1159 |
+
"rewards/margins": 5.432239055633545,
|
| 1160 |
+
"rewards/rejected": -11.558727264404297,
|
| 1161 |
+
"step": 640
|
| 1162 |
+
},
|
| 1163 |
+
{
|
| 1164 |
+
"epoch": 1.0188346085932902,
|
| 1165 |
+
"grad_norm": 4.93217134475708,
|
| 1166 |
+
"learning_rate": 3.019779227044398e-06,
|
| 1167 |
+
"logits/chosen": 13.347890853881836,
|
| 1168 |
+
"logits/rejected": 12.535311698913574,
|
| 1169 |
+
"logps/chosen": -4.311320781707764,
|
| 1170 |
+
"logps/rejected": -7.402029514312744,
|
| 1171 |
+
"loss": 0.2353,
|
| 1172 |
+
"rewards/accuracies": 0.9375,
|
| 1173 |
+
"rewards/chosen": -6.466981410980225,
|
| 1174 |
+
"rewards/margins": 4.636063098907471,
|
| 1175 |
+
"rewards/rejected": -11.103044509887695,
|
| 1176 |
+
"step": 650
|
| 1177 |
+
},
|
| 1178 |
+
{
|
| 1179 |
+
"epoch": 1.0188346085932902,
|
| 1180 |
+
"eval_logits/chosen": 13.296299934387207,
|
| 1181 |
+
"eval_logits/rejected": 12.358522415161133,
|
| 1182 |
+
"eval_logps/chosen": -4.177649974822998,
|
| 1183 |
+
"eval_logps/rejected": -8.009714126586914,
|
| 1184 |
+
"eval_loss": 0.28009530901908875,
|
| 1185 |
+
"eval_rewards/accuracies": 0.9615384340286255,
|
| 1186 |
+
"eval_rewards/chosen": -6.266475200653076,
|
| 1187 |
+
"eval_rewards/margins": 5.748095989227295,
|
| 1188 |
+
"eval_rewards/rejected": -12.014572143554688,
|
| 1189 |
+
"eval_runtime": 14.6091,
|
| 1190 |
+
"eval_samples_per_second": 28.202,
|
| 1191 |
+
"eval_steps_per_second": 3.559,
|
| 1192 |
+
"step": 650
|
| 1193 |
+
},
|
| 1194 |
+
{
|
| 1195 |
+
"epoch": 1.0345301157543654,
|
| 1196 |
+
"grad_norm": 2.0475847721099854,
|
| 1197 |
+
"learning_rate": 2.9684532864643123e-06,
|
| 1198 |
+
"logits/chosen": 13.43058967590332,
|
| 1199 |
+
"logits/rejected": 12.661652565002441,
|
| 1200 |
+
"logps/chosen": -3.9617011547088623,
|
| 1201 |
+
"logps/rejected": -7.092957973480225,
|
| 1202 |
+
"loss": 0.2562,
|
| 1203 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 1204 |
+
"rewards/chosen": -5.942551136016846,
|
| 1205 |
+
"rewards/margins": 4.69688606262207,
|
| 1206 |
+
"rewards/rejected": -10.639436721801758,
|
| 1207 |
+
"step": 660
|
| 1208 |
+
},
|
| 1209 |
+
{
|
| 1210 |
+
"epoch": 1.0502256229154405,
|
| 1211 |
+
"grad_norm": 4.578652381896973,
|
| 1212 |
+
"learning_rate": 2.9169218667902562e-06,
|
| 1213 |
+
"logits/chosen": 13.032757759094238,
|
| 1214 |
+
"logits/rejected": 12.292248725891113,
|
| 1215 |
+
"logps/chosen": -3.679547071456909,
|
| 1216 |
+
"logps/rejected": -7.649219512939453,
|
| 1217 |
+
"loss": 0.2271,
|
| 1218 |
+
"rewards/accuracies": 0.987500011920929,
|
| 1219 |
+
"rewards/chosen": -5.519320487976074,
|
| 1220 |
+
"rewards/margins": 5.9545087814331055,
|
| 1221 |
+
"rewards/rejected": -11.47382926940918,
|
| 1222 |
+
"step": 670
|
| 1223 |
+
},
|
| 1224 |
+
{
|
| 1225 |
+
"epoch": 1.0659211300765157,
|
| 1226 |
+
"grad_norm": 3.0983850955963135,
|
| 1227 |
+
"learning_rate": 2.8652075714060296e-06,
|
| 1228 |
+
"logits/chosen": 12.6952486038208,
|
| 1229 |
+
"logits/rejected": 11.917645454406738,
|
| 1230 |
+
"logps/chosen": -4.468019008636475,
|
| 1231 |
+
"logps/rejected": -8.505170822143555,
|
| 1232 |
+
"loss": 0.179,
|
| 1233 |
+
"rewards/accuracies": 0.9375,
|
| 1234 |
+
"rewards/chosen": -6.702028751373291,
|
| 1235 |
+
"rewards/margins": 6.055727958679199,
|
| 1236 |
+
"rewards/rejected": -12.757757186889648,
|
| 1237 |
+
"step": 680
|
| 1238 |
+
},
|
| 1239 |
+
{
|
| 1240 |
+
"epoch": 1.0816166372375908,
|
| 1241 |
+
"grad_norm": 3.071026086807251,
|
| 1242 |
+
"learning_rate": 2.813333083910761e-06,
|
| 1243 |
+
"logits/chosen": 12.464409828186035,
|
| 1244 |
+
"logits/rejected": 11.680208206176758,
|
| 1245 |
+
"logps/chosen": -4.6339874267578125,
|
| 1246 |
+
"logps/rejected": -8.639189720153809,
|
| 1247 |
+
"loss": 0.2541,
|
| 1248 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 1249 |
+
"rewards/chosen": -6.950981140136719,
|
| 1250 |
+
"rewards/margins": 6.007803440093994,
|
| 1251 |
+
"rewards/rejected": -12.958784103393555,
|
| 1252 |
+
"step": 690
|
| 1253 |
+
},
|
| 1254 |
+
{
|
| 1255 |
+
"epoch": 1.097312144398666,
|
| 1256 |
+
"grad_norm": 1.3732093572616577,
|
| 1257 |
+
"learning_rate": 2.761321158169134e-06,
|
| 1258 |
+
"logits/chosen": 12.308462142944336,
|
| 1259 |
+
"logits/rejected": 11.674012184143066,
|
| 1260 |
+
"logps/chosen": -4.638761043548584,
|
| 1261 |
+
"logps/rejected": -8.936636924743652,
|
| 1262 |
+
"loss": 0.1919,
|
| 1263 |
+
"rewards/accuracies": 0.987500011920929,
|
| 1264 |
+
"rewards/chosen": -6.958141326904297,
|
| 1265 |
+
"rewards/margins": 6.446814060211182,
|
| 1266 |
+
"rewards/rejected": -13.40495491027832,
|
| 1267 |
+
"step": 700
|
| 1268 |
+
},
|
| 1269 |
+
{
|
| 1270 |
+
"epoch": 1.097312144398666,
|
| 1271 |
+
"eval_logits/chosen": 13.217435836791992,
|
| 1272 |
+
"eval_logits/rejected": 12.272343635559082,
|
| 1273 |
+
"eval_logps/chosen": -4.347932815551758,
|
| 1274 |
+
"eval_logps/rejected": -8.441193580627441,
|
| 1275 |
+
"eval_loss": 0.25916898250579834,
|
| 1276 |
+
"eval_rewards/accuracies": 0.9615384340286255,
|
| 1277 |
+
"eval_rewards/chosen": -6.521899223327637,
|
| 1278 |
+
"eval_rewards/margins": 6.139890670776367,
|
| 1279 |
+
"eval_rewards/rejected": -12.661789894104004,
|
| 1280 |
+
"eval_runtime": 14.6102,
|
| 1281 |
+
"eval_samples_per_second": 28.2,
|
| 1282 |
+
"eval_steps_per_second": 3.559,
|
| 1283 |
+
"step": 700
|
| 1284 |
+
},
|
| 1285 |
+
{
|
| 1286 |
+
"epoch": 1.113007651559741,
|
| 1287 |
+
"grad_norm": 3.5739834308624268,
|
| 1288 |
+
"learning_rate": 2.70919460833079e-06,
|
| 1289 |
+
"logits/chosen": 12.828997611999512,
|
| 1290 |
+
"logits/rejected": 12.09227180480957,
|
| 1291 |
+
"logps/chosen": -4.3685622215271,
|
| 1292 |
+
"logps/rejected": -8.672820091247559,
|
| 1293 |
+
"loss": 0.2187,
|
| 1294 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 1295 |
+
"rewards/chosen": -6.552844047546387,
|
| 1296 |
+
"rewards/margins": 6.456387519836426,
|
| 1297 |
+
"rewards/rejected": -13.009231567382812,
|
| 1298 |
+
"step": 710
|
| 1299 |
+
},
|
| 1300 |
+
{
|
| 1301 |
+
"epoch": 1.1287031587208163,
|
| 1302 |
+
"grad_norm": 11.475191116333008,
|
| 1303 |
+
"learning_rate": 2.6569762988232838e-06,
|
| 1304 |
+
"logits/chosen": 13.344253540039062,
|
| 1305 |
+
"logits/rejected": 12.190181732177734,
|
| 1306 |
+
"logps/chosen": -4.567076206207275,
|
| 1307 |
+
"logps/rejected": -8.719629287719727,
|
| 1308 |
+
"loss": 0.1847,
|
| 1309 |
+
"rewards/accuracies": 0.9375,
|
| 1310 |
+
"rewards/chosen": -6.85061502456665,
|
| 1311 |
+
"rewards/margins": 6.228829383850098,
|
| 1312 |
+
"rewards/rejected": -13.079442977905273,
|
| 1313 |
+
"step": 720
|
| 1314 |
+
},
|
| 1315 |
+
{
|
| 1316 |
+
"epoch": 1.1443986658818912,
|
| 1317 |
+
"grad_norm": 1.8145791292190552,
|
| 1318 |
+
"learning_rate": 2.604689134322999e-06,
|
| 1319 |
+
"logits/chosen": 12.794706344604492,
|
| 1320 |
+
"logits/rejected": 12.407793045043945,
|
| 1321 |
+
"logps/chosen": -4.447361946105957,
|
| 1322 |
+
"logps/rejected": -8.19226360321045,
|
| 1323 |
+
"loss": 0.1914,
|
| 1324 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 1325 |
+
"rewards/chosen": -6.67104434967041,
|
| 1326 |
+
"rewards/margins": 5.6173505783081055,
|
| 1327 |
+
"rewards/rejected": -12.288395881652832,
|
| 1328 |
+
"step": 730
|
| 1329 |
+
},
|
| 1330 |
+
{
|
| 1331 |
+
"epoch": 1.1600941730429664,
|
| 1332 |
+
"grad_norm": 3.763374090194702,
|
| 1333 |
+
"learning_rate": 2.5523560497083927e-06,
|
| 1334 |
+
"logits/chosen": 12.97741413116455,
|
| 1335 |
+
"logits/rejected": 12.224584579467773,
|
| 1336 |
+
"logps/chosen": -4.449431419372559,
|
| 1337 |
+
"logps/rejected": -8.755613327026367,
|
| 1338 |
+
"loss": 0.2055,
|
| 1339 |
+
"rewards/accuracies": 0.9375,
|
| 1340 |
+
"rewards/chosen": -6.6741461753845215,
|
| 1341 |
+
"rewards/margins": 6.459272861480713,
|
| 1342 |
+
"rewards/rejected": -13.133418083190918,
|
| 1343 |
+
"step": 740
|
| 1344 |
+
},
|
| 1345 |
+
{
|
| 1346 |
+
"epoch": 1.1757896802040415,
|
| 1347 |
+
"grad_norm": 2.816288948059082,
|
| 1348 |
+
"learning_rate": 2.5e-06,
|
| 1349 |
+
"logits/chosen": 12.736323356628418,
|
| 1350 |
+
"logits/rejected": 12.039944648742676,
|
| 1351 |
+
"logps/chosen": -4.300374507904053,
|
| 1352 |
+
"logps/rejected": -8.570680618286133,
|
| 1353 |
+
"loss": 0.185,
|
| 1354 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1355 |
+
"rewards/chosen": -6.4505615234375,
|
| 1356 |
+
"rewards/margins": 6.405457973480225,
|
| 1357 |
+
"rewards/rejected": -12.856019973754883,
|
| 1358 |
+
"step": 750
|
| 1359 |
+
},
|
| 1360 |
+
{
|
| 1361 |
+
"epoch": 1.1757896802040415,
|
| 1362 |
+
"eval_logits/chosen": 13.174198150634766,
|
| 1363 |
+
"eval_logits/rejected": 12.189348220825195,
|
| 1364 |
+
"eval_logps/chosen": -4.302711009979248,
|
| 1365 |
+
"eval_logps/rejected": -8.745194435119629,
|
| 1366 |
+
"eval_loss": 0.24307754635810852,
|
| 1367 |
+
"eval_rewards/accuracies": 0.9615384340286255,
|
| 1368 |
+
"eval_rewards/chosen": -6.454067707061768,
|
| 1369 |
+
"eval_rewards/margins": 6.663723945617676,
|
| 1370 |
+
"eval_rewards/rejected": -13.117791175842285,
|
| 1371 |
+
"eval_runtime": 14.5945,
|
| 1372 |
+
"eval_samples_per_second": 28.23,
|
| 1373 |
+
"eval_steps_per_second": 3.563,
|
| 1374 |
+
"step": 750
|
| 1375 |
+
},
|
| 1376 |
+
{
|
| 1377 |
+
"epoch": 1.1914851873651167,
|
| 1378 |
+
"grad_norm": 2.6783711910247803,
|
| 1379 |
+
"learning_rate": 2.447643950291608e-06,
|
| 1380 |
+
"logits/chosen": 12.992940902709961,
|
| 1381 |
+
"logits/rejected": 12.16160774230957,
|
| 1382 |
+
"logps/chosen": -4.402331829071045,
|
| 1383 |
+
"logps/rejected": -8.848995208740234,
|
| 1384 |
+
"loss": 0.1958,
|
| 1385 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 1386 |
+
"rewards/chosen": -6.603497505187988,
|
| 1387 |
+
"rewards/margins": 6.6699957847595215,
|
| 1388 |
+
"rewards/rejected": -13.273493766784668,
|
| 1389 |
+
"step": 760
|
| 1390 |
+
},
|
| 1391 |
+
{
|
| 1392 |
+
"epoch": 1.2071806945261918,
|
| 1393 |
+
"grad_norm": 1.580244779586792,
|
| 1394 |
+
"learning_rate": 2.3953108656770018e-06,
|
| 1395 |
+
"logits/chosen": 12.522805213928223,
|
| 1396 |
+
"logits/rejected": 11.664339065551758,
|
| 1397 |
+
"logps/chosen": -3.9605941772460938,
|
| 1398 |
+
"logps/rejected": -8.103001594543457,
|
| 1399 |
+
"loss": 0.1921,
|
| 1400 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1401 |
+
"rewards/chosen": -5.940891265869141,
|
| 1402 |
+
"rewards/margins": 6.213610649108887,
|
| 1403 |
+
"rewards/rejected": -12.154502868652344,
|
| 1404 |
+
"step": 770
|
| 1405 |
+
},
|
| 1406 |
+
{
|
| 1407 |
+
"epoch": 1.222876201687267,
|
| 1408 |
+
"grad_norm": 2.2647345066070557,
|
| 1409 |
+
"learning_rate": 2.3430237011767166e-06,
|
| 1410 |
+
"logits/chosen": 12.654630661010742,
|
| 1411 |
+
"logits/rejected": 11.677509307861328,
|
| 1412 |
+
"logps/chosen": -4.634467601776123,
|
| 1413 |
+
"logps/rejected": -9.647830963134766,
|
| 1414 |
+
"loss": 0.1731,
|
| 1415 |
+
"rewards/accuracies": 0.987500011920929,
|
| 1416 |
+
"rewards/chosen": -6.951702117919922,
|
| 1417 |
+
"rewards/margins": 7.520044803619385,
|
| 1418 |
+
"rewards/rejected": -14.471745491027832,
|
| 1419 |
+
"step": 780
|
| 1420 |
+
},
|
| 1421 |
+
{
|
| 1422 |
+
"epoch": 1.2385717088483421,
|
| 1423 |
+
"grad_norm": 3.8534202575683594,
|
| 1424 |
+
"learning_rate": 2.290805391669212e-06,
|
| 1425 |
+
"logits/chosen": 12.548233032226562,
|
| 1426 |
+
"logits/rejected": 11.633279800415039,
|
| 1427 |
+
"logps/chosen": -4.690009593963623,
|
| 1428 |
+
"logps/rejected": -8.73483657836914,
|
| 1429 |
+
"loss": 0.1844,
|
| 1430 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1431 |
+
"rewards/chosen": -7.0350141525268555,
|
| 1432 |
+
"rewards/margins": 6.067240238189697,
|
| 1433 |
+
"rewards/rejected": -13.102254867553711,
|
| 1434 |
+
"step": 790
|
| 1435 |
+
},
|
| 1436 |
+
{
|
| 1437 |
+
"epoch": 1.2542672160094173,
|
| 1438 |
+
"grad_norm": 2.8962628841400146,
|
| 1439 |
+
"learning_rate": 2.238678841830867e-06,
|
| 1440 |
+
"logits/chosen": 11.90459156036377,
|
| 1441 |
+
"logits/rejected": 11.352374076843262,
|
| 1442 |
+
"logps/chosen": -4.448246479034424,
|
| 1443 |
+
"logps/rejected": -8.658299446105957,
|
| 1444 |
+
"loss": 0.1638,
|
| 1445 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1446 |
+
"rewards/chosen": -6.67236852645874,
|
| 1447 |
+
"rewards/margins": 6.315079689025879,
|
| 1448 |
+
"rewards/rejected": -12.987447738647461,
|
| 1449 |
+
"step": 800
|
| 1450 |
+
},
|
| 1451 |
+
{
|
| 1452 |
+
"epoch": 1.2542672160094173,
|
| 1453 |
+
"eval_logits/chosen": 12.893868446350098,
|
| 1454 |
+
"eval_logits/rejected": 11.863032341003418,
|
| 1455 |
+
"eval_logps/chosen": -4.604159832000732,
|
| 1456 |
+
"eval_logps/rejected": -9.257932662963867,
|
| 1457 |
+
"eval_loss": 0.21881447732448578,
|
| 1458 |
+
"eval_rewards/accuracies": 0.9615384340286255,
|
| 1459 |
+
"eval_rewards/chosen": -6.906240463256836,
|
| 1460 |
+
"eval_rewards/margins": 6.980658531188965,
|
| 1461 |
+
"eval_rewards/rejected": -13.886898040771484,
|
| 1462 |
+
"eval_runtime": 14.6079,
|
| 1463 |
+
"eval_samples_per_second": 28.204,
|
| 1464 |
+
"eval_steps_per_second": 3.56,
|
| 1465 |
+
"step": 800
|
| 1466 |
+
},
|
| 1467 |
+
{
|
| 1468 |
+
"epoch": 1.2699627231704924,
|
| 1469 |
+
"grad_norm": 3.5829524993896484,
|
| 1470 |
+
"learning_rate": 2.186666916089239e-06,
|
| 1471 |
+
"logits/chosen": 12.411908149719238,
|
| 1472 |
+
"logits/rejected": 11.730254173278809,
|
| 1473 |
+
"logps/chosen": -4.65117883682251,
|
| 1474 |
+
"logps/rejected": -9.094849586486816,
|
| 1475 |
+
"loss": 0.1827,
|
| 1476 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 1477 |
+
"rewards/chosen": -6.976768493652344,
|
| 1478 |
+
"rewards/margins": 6.665505886077881,
|
| 1479 |
+
"rewards/rejected": -13.642274856567383,
|
| 1480 |
+
"step": 810
|
| 1481 |
+
},
|
| 1482 |
+
{
|
| 1483 |
+
"epoch": 1.2856582303315676,
|
| 1484 |
+
"grad_norm": 2.836439371109009,
|
| 1485 |
+
"learning_rate": 2.134792428593971e-06,
|
| 1486 |
+
"logits/chosen": 12.173138618469238,
|
| 1487 |
+
"logits/rejected": 11.570637702941895,
|
| 1488 |
+
"logps/chosen": -4.847201824188232,
|
| 1489 |
+
"logps/rejected": -8.849002838134766,
|
| 1490 |
+
"loss": 0.1716,
|
| 1491 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1492 |
+
"rewards/chosen": -7.270802974700928,
|
| 1493 |
+
"rewards/margins": 6.002701282501221,
|
| 1494 |
+
"rewards/rejected": -13.273503303527832,
|
| 1495 |
+
"step": 820
|
| 1496 |
+
},
|
| 1497 |
+
{
|
| 1498 |
+
"epoch": 1.3013537374926427,
|
| 1499 |
+
"grad_norm": 3.5850038528442383,
|
| 1500 |
+
"learning_rate": 2.0830781332097446e-06,
|
| 1501 |
+
"logits/chosen": 12.883340835571289,
|
| 1502 |
+
"logits/rejected": 12.123892784118652,
|
| 1503 |
+
"logps/chosen": -4.7816948890686035,
|
| 1504 |
+
"logps/rejected": -9.729695320129395,
|
| 1505 |
+
"loss": 0.1652,
|
| 1506 |
+
"rewards/accuracies": 0.9375,
|
| 1507 |
+
"rewards/chosen": -7.172542572021484,
|
| 1508 |
+
"rewards/margins": 7.421999931335449,
|
| 1509 |
+
"rewards/rejected": -14.594541549682617,
|
| 1510 |
+
"step": 830
|
| 1511 |
+
},
|
| 1512 |
+
{
|
| 1513 |
+
"epoch": 1.3170492446537179,
|
| 1514 |
+
"grad_norm": 1.9497729539871216,
|
| 1515 |
+
"learning_rate": 2.031546713535688e-06,
|
| 1516 |
+
"logits/chosen": 12.579580307006836,
|
| 1517 |
+
"logits/rejected": 11.674839973449707,
|
| 1518 |
+
"logps/chosen": -4.67056131362915,
|
| 1519 |
+
"logps/rejected": -9.12755298614502,
|
| 1520 |
+
"loss": 0.1881,
|
| 1521 |
+
"rewards/accuracies": 0.9375,
|
| 1522 |
+
"rewards/chosen": -7.0058417320251465,
|
| 1523 |
+
"rewards/margins": 6.68548583984375,
|
| 1524 |
+
"rewards/rejected": -13.691328048706055,
|
| 1525 |
+
"step": 840
|
| 1526 |
+
},
|
| 1527 |
+
{
|
| 1528 |
+
"epoch": 1.332744751814793,
|
| 1529 |
+
"grad_norm": 2.5810794830322266,
|
| 1530 |
+
"learning_rate": 1.9802207729556023e-06,
|
| 1531 |
+
"logits/chosen": 12.369009017944336,
|
| 1532 |
+
"logits/rejected": 11.520233154296875,
|
| 1533 |
+
"logps/chosen": -4.780073165893555,
|
| 1534 |
+
"logps/rejected": -9.808363914489746,
|
| 1535 |
+
"loss": 0.2051,
|
| 1536 |
+
"rewards/accuracies": 0.987500011920929,
|
| 1537 |
+
"rewards/chosen": -7.17011022567749,
|
| 1538 |
+
"rewards/margins": 7.542436122894287,
|
| 1539 |
+
"rewards/rejected": -14.712547302246094,
|
| 1540 |
+
"step": 850
|
| 1541 |
+
},
|
| 1542 |
+
{
|
| 1543 |
+
"epoch": 1.332744751814793,
|
| 1544 |
+
"eval_logits/chosen": 12.70075511932373,
|
| 1545 |
+
"eval_logits/rejected": 11.632950782775879,
|
| 1546 |
+
"eval_logps/chosen": -4.634037494659424,
|
| 1547 |
+
"eval_logps/rejected": -9.374123573303223,
|
| 1548 |
+
"eval_loss": 0.20189520716667175,
|
| 1549 |
+
"eval_rewards/accuracies": 0.9615384340286255,
|
| 1550 |
+
"eval_rewards/chosen": -6.951055526733398,
|
| 1551 |
+
"eval_rewards/margins": 7.110130310058594,
|
| 1552 |
+
"eval_rewards/rejected": -14.061185836791992,
|
| 1553 |
+
"eval_runtime": 14.6066,
|
| 1554 |
+
"eval_samples_per_second": 28.206,
|
| 1555 |
+
"eval_steps_per_second": 3.56,
|
| 1556 |
+
"step": 850
|
| 1557 |
+
},
|
| 1558 |
+
{
|
| 1559 |
+
"epoch": 1.3484402589758682,
|
| 1560 |
+
"grad_norm": 5.187539577484131,
|
| 1561 |
+
"learning_rate": 1.9291228247233607e-06,
|
| 1562 |
+
"logits/chosen": 12.224112510681152,
|
| 1563 |
+
"logits/rejected": 11.01191234588623,
|
| 1564 |
+
"logps/chosen": -4.748395919799805,
|
| 1565 |
+
"logps/rejected": -9.206069946289062,
|
| 1566 |
+
"loss": 0.2083,
|
| 1567 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1568 |
+
"rewards/chosen": -7.122593879699707,
|
| 1569 |
+
"rewards/margins": 6.686511039733887,
|
| 1570 |
+
"rewards/rejected": -13.809106826782227,
|
| 1571 |
+
"step": 860
|
| 1572 |
+
},
|
| 1573 |
+
{
|
| 1574 |
+
"epoch": 1.3641357661369433,
|
| 1575 |
+
"grad_norm": 1.1569608449935913,
|
| 1576 |
+
"learning_rate": 1.8782752820878636e-06,
|
| 1577 |
+
"logits/chosen": 11.579218864440918,
|
| 1578 |
+
"logits/rejected": 10.995697975158691,
|
| 1579 |
+
"logps/chosen": -4.387156009674072,
|
| 1580 |
+
"logps/rejected": -8.937141418457031,
|
| 1581 |
+
"loss": 0.1782,
|
| 1582 |
+
"rewards/accuracies": 0.987500011920929,
|
| 1583 |
+
"rewards/chosen": -6.580733299255371,
|
| 1584 |
+
"rewards/margins": 6.824978828430176,
|
| 1585 |
+
"rewards/rejected": -13.405713081359863,
|
| 1586 |
+
"step": 870
|
| 1587 |
+
},
|
| 1588 |
+
{
|
| 1589 |
+
"epoch": 1.3798312732980185,
|
| 1590 |
+
"grad_norm": 5.9647111892700195,
|
| 1591 |
+
"learning_rate": 1.827700448461836e-06,
|
| 1592 |
+
"logits/chosen": 12.533722877502441,
|
| 1593 |
+
"logits/rejected": 12.121667861938477,
|
| 1594 |
+
"logps/chosen": -3.9464848041534424,
|
| 1595 |
+
"logps/rejected": -7.782477378845215,
|
| 1596 |
+
"loss": 0.237,
|
| 1597 |
+
"rewards/accuracies": 0.925000011920929,
|
| 1598 |
+
"rewards/chosen": -5.919726848602295,
|
| 1599 |
+
"rewards/margins": 5.753990173339844,
|
| 1600 |
+
"rewards/rejected": -11.67371654510498,
|
| 1601 |
+
"step": 880
|
| 1602 |
+
},
|
| 1603 |
+
{
|
| 1604 |
+
"epoch": 1.3955267804590936,
|
| 1605 |
+
"grad_norm": 3.4284207820892334,
|
| 1606 |
+
"learning_rate": 1.7774205076388207e-06,
|
| 1607 |
+
"logits/chosen": 12.185909271240234,
|
| 1608 |
+
"logits/rejected": 11.781715393066406,
|
| 1609 |
+
"logps/chosen": -4.110888481140137,
|
| 1610 |
+
"logps/rejected": -8.276056289672852,
|
| 1611 |
+
"loss": 0.1904,
|
| 1612 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 1613 |
+
"rewards/chosen": -6.166332721710205,
|
| 1614 |
+
"rewards/margins": 6.2477498054504395,
|
| 1615 |
+
"rewards/rejected": -12.414082527160645,
|
| 1616 |
+
"step": 890
|
| 1617 |
+
},
|
| 1618 |
+
{
|
| 1619 |
+
"epoch": 1.4112222876201688,
|
| 1620 |
+
"grad_norm": 2.537830352783203,
|
| 1621 |
+
"learning_rate": 1.7274575140626318e-06,
|
| 1622 |
+
"logits/chosen": 12.28978443145752,
|
| 1623 |
+
"logits/rejected": 11.305529594421387,
|
| 1624 |
+
"logps/chosen": -3.9210026264190674,
|
| 1625 |
+
"logps/rejected": -8.545848846435547,
|
| 1626 |
+
"loss": 0.1508,
|
| 1627 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 1628 |
+
"rewards/chosen": -5.881503582000732,
|
| 1629 |
+
"rewards/margins": 6.937270164489746,
|
| 1630 |
+
"rewards/rejected": -12.818774223327637,
|
| 1631 |
+
"step": 900
|
| 1632 |
+
},
|
| 1633 |
+
{
|
| 1634 |
+
"epoch": 1.4112222876201688,
|
| 1635 |
+
"eval_logits/chosen": 12.542530059814453,
|
| 1636 |
+
"eval_logits/rejected": 11.450419425964355,
|
| 1637 |
+
"eval_logps/chosen": -4.722529888153076,
|
| 1638 |
+
"eval_logps/rejected": -9.546163558959961,
|
| 1639 |
+
"eval_loss": 0.18640650808811188,
|
| 1640 |
+
"eval_rewards/accuracies": 0.9807692170143127,
|
| 1641 |
+
"eval_rewards/chosen": -7.083794593811035,
|
| 1642 |
+
"eval_rewards/margins": 7.235450744628906,
|
| 1643 |
+
"eval_rewards/rejected": -14.319246292114258,
|
| 1644 |
+
"eval_runtime": 14.6107,
|
| 1645 |
+
"eval_samples_per_second": 28.199,
|
| 1646 |
+
"eval_steps_per_second": 3.559,
|
| 1647 |
+
"step": 900
|
| 1648 |
+
},
|
| 1649 |
+
{
|
| 1650 |
+
"epoch": 1.426917794781244,
|
| 1651 |
+
"grad_norm": 1.1588759422302246,
|
| 1652 |
+
"learning_rate": 1.677833383153542e-06,
|
| 1653 |
+
"logits/chosen": 12.432401657104492,
|
| 1654 |
+
"logits/rejected": 11.247361183166504,
|
| 1655 |
+
"logps/chosen": -4.893294334411621,
|
| 1656 |
+
"logps/rejected": -9.311574935913086,
|
| 1657 |
+
"loss": 0.1446,
|
| 1658 |
+
"rewards/accuracies": 0.987500011920929,
|
| 1659 |
+
"rewards/chosen": -7.339941501617432,
|
| 1660 |
+
"rewards/margins": 6.627419948577881,
|
| 1661 |
+
"rewards/rejected": -13.967363357543945,
|
| 1662 |
+
"step": 910
|
| 1663 |
+
},
|
| 1664 |
+
{
|
| 1665 |
+
"epoch": 1.442613301942319,
|
| 1666 |
+
"grad_norm": 2.4712307453155518,
|
| 1667 |
+
"learning_rate": 1.6285698816954626e-06,
|
| 1668 |
+
"logits/chosen": 12.139798164367676,
|
| 1669 |
+
"logits/rejected": 11.248186111450195,
|
| 1670 |
+
"logps/chosen": -5.167681694030762,
|
| 1671 |
+
"logps/rejected": -10.253265380859375,
|
| 1672 |
+
"loss": 0.163,
|
| 1673 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 1674 |
+
"rewards/chosen": -7.751523017883301,
|
| 1675 |
+
"rewards/margins": 7.6283769607543945,
|
| 1676 |
+
"rewards/rejected": -15.379899978637695,
|
| 1677 |
+
"step": 920
|
| 1678 |
+
},
|
| 1679 |
+
{
|
| 1680 |
+
"epoch": 1.4583088091033942,
|
| 1681 |
+
"grad_norm": 1.881538987159729,
|
| 1682 |
+
"learning_rate": 1.5796886182883053e-06,
|
| 1683 |
+
"logits/chosen": 11.9426908493042,
|
| 1684 |
+
"logits/rejected": 11.341384887695312,
|
| 1685 |
+
"logps/chosen": -5.030823230743408,
|
| 1686 |
+
"logps/rejected": -9.114843368530273,
|
| 1687 |
+
"loss": 0.1773,
|
| 1688 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1689 |
+
"rewards/chosen": -7.54623556137085,
|
| 1690 |
+
"rewards/margins": 6.126028060913086,
|
| 1691 |
+
"rewards/rejected": -13.672262191772461,
|
| 1692 |
+
"step": 930
|
| 1693 |
+
},
|
| 1694 |
+
{
|
| 1695 |
+
"epoch": 1.4740043162644694,
|
| 1696 |
+
"grad_norm": 5.751014709472656,
|
| 1697 |
+
"learning_rate": 1.5312110338697427e-06,
|
| 1698 |
+
"logits/chosen": 11.921320915222168,
|
| 1699 |
+
"logits/rejected": 11.005578994750977,
|
| 1700 |
+
"logps/chosen": -4.730319976806641,
|
| 1701 |
+
"logps/rejected": -9.198332786560059,
|
| 1702 |
+
"loss": 0.2058,
|
| 1703 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 1704 |
+
"rewards/chosen": -7.095480442047119,
|
| 1705 |
+
"rewards/margins": 6.702019691467285,
|
| 1706 |
+
"rewards/rejected": -13.797500610351562,
|
| 1707 |
+
"step": 940
|
| 1708 |
+
},
|
| 1709 |
+
{
|
| 1710 |
+
"epoch": 1.4896998234255445,
|
| 1711 |
+
"grad_norm": 2.656139850616455,
|
| 1712 |
+
"learning_rate": 1.4831583923105e-06,
|
| 1713 |
+
"logits/chosen": 12.132013320922852,
|
| 1714 |
+
"logits/rejected": 11.254085540771484,
|
| 1715 |
+
"logps/chosen": -4.832255840301514,
|
| 1716 |
+
"logps/rejected": -9.473350524902344,
|
| 1717 |
+
"loss": 0.1827,
|
| 1718 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 1719 |
+
"rewards/chosen": -7.24838399887085,
|
| 1720 |
+
"rewards/margins": 6.961642265319824,
|
| 1721 |
+
"rewards/rejected": -14.210027694702148,
|
| 1722 |
+
"step": 950
|
| 1723 |
+
},
|
| 1724 |
+
{
|
| 1725 |
+
"epoch": 1.4896998234255445,
|
| 1726 |
+
"eval_logits/chosen": 12.534186363220215,
|
| 1727 |
+
"eval_logits/rejected": 11.43524169921875,
|
| 1728 |
+
"eval_logps/chosen": -4.488946914672852,
|
| 1729 |
+
"eval_logps/rejected": -9.457845687866211,
|
| 1730 |
+
"eval_loss": 0.1744592934846878,
|
| 1731 |
+
"eval_rewards/accuracies": 0.9807692170143127,
|
| 1732 |
+
"eval_rewards/chosen": -6.733420372009277,
|
| 1733 |
+
"eval_rewards/margins": 7.4533467292785645,
|
| 1734 |
+
"eval_rewards/rejected": -14.186767578125,
|
| 1735 |
+
"eval_runtime": 14.6075,
|
| 1736 |
+
"eval_samples_per_second": 28.205,
|
| 1737 |
+
"eval_steps_per_second": 3.56,
|
| 1738 |
+
"step": 950
|
| 1739 |
+
},
|
| 1740 |
+
{
|
| 1741 |
+
"epoch": 1.5053953305866195,
|
| 1742 |
+
"grad_norm": 4.976431369781494,
|
| 1743 |
+
"learning_rate": 1.4355517710873184e-06,
|
| 1744 |
+
"logits/chosen": 12.053160667419434,
|
| 1745 |
+
"logits/rejected": 11.461877822875977,
|
| 1746 |
+
"logps/chosen": -4.3384108543396,
|
| 1747 |
+
"logps/rejected": -8.933584213256836,
|
| 1748 |
+
"loss": 0.1559,
|
| 1749 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1750 |
+
"rewards/chosen": -6.5076165199279785,
|
| 1751 |
+
"rewards/margins": 6.892758846282959,
|
| 1752 |
+
"rewards/rejected": -13.400375366210938,
|
| 1753 |
+
"step": 960
|
| 1754 |
+
},
|
| 1755 |
+
{
|
| 1756 |
+
"epoch": 1.5210908377476948,
|
| 1757 |
+
"grad_norm": 2.786695718765259,
|
| 1758 |
+
"learning_rate": 1.388412052037682e-06,
|
| 1759 |
+
"logits/chosen": 12.072885513305664,
|
| 1760 |
+
"logits/rejected": 10.973172187805176,
|
| 1761 |
+
"logps/chosen": -4.89219331741333,
|
| 1762 |
+
"logps/rejected": -9.545186996459961,
|
| 1763 |
+
"loss": 0.1334,
|
| 1764 |
+
"rewards/accuracies": 0.9375,
|
| 1765 |
+
"rewards/chosen": -7.338289737701416,
|
| 1766 |
+
"rewards/margins": 6.979489803314209,
|
| 1767 |
+
"rewards/rejected": -14.317779541015625,
|
| 1768 |
+
"step": 970
|
| 1769 |
+
},
|
| 1770 |
+
{
|
| 1771 |
+
"epoch": 1.5367863449087698,
|
| 1772 |
+
"grad_norm": 5.252014636993408,
|
| 1773 |
+
"learning_rate": 1.3417599122003464e-06,
|
| 1774 |
+
"logits/chosen": 12.37597370147705,
|
| 1775 |
+
"logits/rejected": 11.43901538848877,
|
| 1776 |
+
"logps/chosen": -5.274794101715088,
|
| 1777 |
+
"logps/rejected": -9.821117401123047,
|
| 1778 |
+
"loss": 0.1477,
|
| 1779 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1780 |
+
"rewards/chosen": -7.912191867828369,
|
| 1781 |
+
"rewards/margins": 6.819484710693359,
|
| 1782 |
+
"rewards/rejected": -14.73167610168457,
|
| 1783 |
+
"step": 980
|
| 1784 |
+
},
|
| 1785 |
+
{
|
| 1786 |
+
"epoch": 1.5524818520698451,
|
| 1787 |
+
"grad_norm": 2.243185520172119,
|
| 1788 |
+
"learning_rate": 1.2956158147457116e-06,
|
| 1789 |
+
"logits/chosen": 11.921972274780273,
|
| 1790 |
+
"logits/rejected": 11.30908489227295,
|
| 1791 |
+
"logps/chosen": -5.393095970153809,
|
| 1792 |
+
"logps/rejected": -10.215553283691406,
|
| 1793 |
+
"loss": 0.1561,
|
| 1794 |
+
"rewards/accuracies": 0.949999988079071,
|
| 1795 |
+
"rewards/chosen": -8.089644432067871,
|
| 1796 |
+
"rewards/margins": 7.233686923980713,
|
| 1797 |
+
"rewards/rejected": -15.323330879211426,
|
| 1798 |
+
"step": 990
|
| 1799 |
+
},
|
| 1800 |
+
{
|
| 1801 |
+
"epoch": 1.56817735923092,
|
| 1802 |
+
"grad_norm": 3.4886062145233154,
|
| 1803 |
+
"learning_rate": 1.2500000000000007e-06,
|
| 1804 |
+
"logits/chosen": 11.753881454467773,
|
| 1805 |
+
"logits/rejected": 11.269731521606445,
|
| 1806 |
+
"logps/chosen": -5.015429496765137,
|
| 1807 |
+
"logps/rejected": -10.169350624084473,
|
| 1808 |
+
"loss": 0.1771,
|
| 1809 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 1810 |
+
"rewards/chosen": -7.523144721984863,
|
| 1811 |
+
"rewards/margins": 7.7308807373046875,
|
| 1812 |
+
"rewards/rejected": -15.25402545928955,
|
| 1813 |
+
"step": 1000
|
| 1814 |
+
},
|
| 1815 |
+
{
|
| 1816 |
+
"epoch": 1.56817735923092,
|
| 1817 |
+
"eval_logits/chosen": 12.437786102294922,
|
| 1818 |
+
"eval_logits/rejected": 11.32490348815918,
|
| 1819 |
+
"eval_logps/chosen": -4.6446380615234375,
|
| 1820 |
+
"eval_logps/rejected": -9.707358360290527,
|
| 1821 |
+
"eval_loss": 0.16595196723937988,
|
| 1822 |
+
"eval_rewards/accuracies": 0.9807692170143127,
|
| 1823 |
+
"eval_rewards/chosen": -6.966956615447998,
|
| 1824 |
+
"eval_rewards/margins": 7.594080448150635,
|
| 1825 |
+
"eval_rewards/rejected": -14.561037063598633,
|
| 1826 |
+
"eval_runtime": 14.6109,
|
| 1827 |
+
"eval_samples_per_second": 28.198,
|
| 1828 |
+
"eval_steps_per_second": 3.559,
|
| 1829 |
+
"step": 1000
|
| 1830 |
+
},
|
| 1831 |
+
{
|
| 1832 |
+
"epoch": 1.5838728663919954,
|
| 1833 |
+
"grad_norm": 1.7607649564743042,
|
| 1834 |
+
"learning_rate": 1.204932476567175e-06,
|
| 1835 |
+
"logits/chosen": 11.258343696594238,
|
| 1836 |
+
"logits/rejected": 10.613565444946289,
|
| 1837 |
+
"logps/chosen": -4.758576393127441,
|
| 1838 |
+
"logps/rejected": -9.250303268432617,
|
| 1839 |
+
"loss": 0.1668,
|
| 1840 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 1841 |
+
"rewards/chosen": -7.137864589691162,
|
| 1842 |
+
"rewards/margins": 6.7375898361206055,
|
| 1843 |
+
"rewards/rejected": -13.875455856323242,
|
| 1844 |
+
"step": 1010
|
| 1845 |
+
},
|
| 1846 |
+
{
|
| 1847 |
+
"epoch": 1.5995683735530704,
|
| 1848 |
+
"grad_norm": 3.2954118251800537,
|
| 1849 |
+
"learning_rate": 1.160433012552508e-06,
|
| 1850 |
+
"logits/chosen": 12.226592063903809,
|
| 1851 |
+
"logits/rejected": 11.355180740356445,
|
| 1852 |
+
"logps/chosen": -4.198899269104004,
|
| 1853 |
+
"logps/rejected": -8.675564765930176,
|
| 1854 |
+
"loss": 0.1824,
|
| 1855 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 1856 |
+
"rewards/chosen": -6.298348903656006,
|
| 1857 |
+
"rewards/margins": 6.714995384216309,
|
| 1858 |
+
"rewards/rejected": -13.013345718383789,
|
| 1859 |
+
"step": 1020
|
| 1860 |
+
},
|
| 1861 |
+
{
|
| 1862 |
+
"epoch": 1.6152638807141457,
|
| 1863 |
+
"grad_norm": 2.833040237426758,
|
| 1864 |
+
"learning_rate": 1.11652112689164e-06,
|
| 1865 |
+
"logits/chosen": 11.937528610229492,
|
| 1866 |
+
"logits/rejected": 11.127592086791992,
|
| 1867 |
+
"logps/chosen": -4.438723087310791,
|
| 1868 |
+
"logps/rejected": -9.106756210327148,
|
| 1869 |
+
"loss": 0.1567,
|
| 1870 |
+
"rewards/accuracies": 0.987500011920929,
|
| 1871 |
+
"rewards/chosen": -6.658083915710449,
|
| 1872 |
+
"rewards/margins": 7.002050876617432,
|
| 1873 |
+
"rewards/rejected": -13.660135269165039,
|
| 1874 |
+
"step": 1030
|
| 1875 |
+
},
|
| 1876 |
+
{
|
| 1877 |
+
"epoch": 1.6309593878752207,
|
| 1878 |
+
"grad_norm": 2.0544891357421875,
|
| 1879 |
+
"learning_rate": 1.073216080788921e-06,
|
| 1880 |
+
"logits/chosen": 12.242444038391113,
|
| 1881 |
+
"logits/rejected": 11.25694465637207,
|
| 1882 |
+
"logps/chosen": -4.525513172149658,
|
| 1883 |
+
"logps/rejected": -9.212661743164062,
|
| 1884 |
+
"loss": 0.1682,
|
| 1885 |
+
"rewards/accuracies": 0.987500011920929,
|
| 1886 |
+
"rewards/chosen": -6.78826904296875,
|
| 1887 |
+
"rewards/margins": 7.030723571777344,
|
| 1888 |
+
"rewards/rejected": -13.818992614746094,
|
| 1889 |
+
"step": 1040
|
| 1890 |
+
},
|
| 1891 |
+
{
|
| 1892 |
+
"epoch": 1.6466548950362958,
|
| 1893 |
+
"grad_norm": 1.6990854740142822,
|
| 1894 |
+
"learning_rate": 1.0305368692688175e-06,
|
| 1895 |
+
"logits/chosen": 11.99101448059082,
|
| 1896 |
+
"logits/rejected": 11.161852836608887,
|
| 1897 |
+
"logps/chosen": -4.299143314361572,
|
| 1898 |
+
"logps/rejected": -9.032526016235352,
|
| 1899 |
+
"loss": 0.1143,
|
| 1900 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 1901 |
+
"rewards/chosen": -6.448714256286621,
|
| 1902 |
+
"rewards/margins": 7.100072383880615,
|
| 1903 |
+
"rewards/rejected": -13.548787117004395,
|
| 1904 |
+
"step": 1050
|
| 1905 |
+
},
|
| 1906 |
+
{
|
| 1907 |
+
"epoch": 1.6466548950362958,
|
| 1908 |
+
"eval_logits/chosen": 12.486811637878418,
|
| 1909 |
+
"eval_logits/rejected": 11.354048728942871,
|
| 1910 |
+
"eval_logps/chosen": -4.389483451843262,
|
| 1911 |
+
"eval_logps/rejected": -9.478667259216309,
|
| 1912 |
+
"eval_loss": 0.16117088496685028,
|
| 1913 |
+
"eval_rewards/accuracies": 0.9807692170143127,
|
| 1914 |
+
"eval_rewards/chosen": -6.584225654602051,
|
| 1915 |
+
"eval_rewards/margins": 7.633774280548096,
|
| 1916 |
+
"eval_rewards/rejected": -14.217998504638672,
|
| 1917 |
+
"eval_runtime": 14.6006,
|
| 1918 |
+
"eval_samples_per_second": 28.218,
|
| 1919 |
+
"eval_steps_per_second": 3.561,
|
| 1920 |
+
"step": 1050
|
| 1921 |
+
},
|
| 1922 |
+
{
|
| 1923 |
+
"epoch": 1.662350402197371,
|
| 1924 |
+
"grad_norm": 3.3203821182250977,
|
| 1925 |
+
"learning_rate": 9.88502212844063e-07,
|
| 1926 |
+
"logits/chosen": 12.17997932434082,
|
| 1927 |
+
"logits/rejected": 11.155837059020996,
|
| 1928 |
+
"logps/chosen": -4.206795692443848,
|
| 1929 |
+
"logps/rejected": -9.341482162475586,
|
| 1930 |
+
"loss": 0.1631,
|
| 1931 |
+
"rewards/accuracies": 1.0,
|
| 1932 |
+
"rewards/chosen": -6.310193061828613,
|
| 1933 |
+
"rewards/margins": 7.702028751373291,
|
| 1934 |
+
"rewards/rejected": -14.012222290039062,
|
| 1935 |
+
"step": 1060
|
| 1936 |
+
},
|
| 1937 |
+
{
|
| 1938 |
+
"epoch": 1.678045909358446,
|
| 1939 |
+
"grad_norm": 2.4238297939300537,
|
| 1940 |
+
"learning_rate": 9.471305493042243e-07,
|
| 1941 |
+
"logits/chosen": 11.849737167358398,
|
| 1942 |
+
"logits/rejected": 10.839106559753418,
|
| 1943 |
+
"logps/chosen": -4.497702598571777,
|
| 1944 |
+
"logps/rejected": -9.431967735290527,
|
| 1945 |
+
"loss": 0.1347,
|
| 1946 |
+
"rewards/accuracies": 1.0,
|
| 1947 |
+
"rewards/chosen": -6.746554374694824,
|
| 1948 |
+
"rewards/margins": 7.401397705078125,
|
| 1949 |
+
"rewards/rejected": -14.14795207977295,
|
| 1950 |
+
"step": 1070
|
| 1951 |
+
},
|
| 1952 |
+
{
|
| 1953 |
+
"epoch": 1.6937414165195213,
|
| 1954 |
+
"grad_norm": 1.895890712738037,
|
| 1955 |
+
"learning_rate": 9.064400256282757e-07,
|
| 1956 |
+
"logits/chosen": 12.01026725769043,
|
| 1957 |
+
"logits/rejected": 11.010753631591797,
|
| 1958 |
+
"logps/chosen": -4.6886305809021,
|
| 1959 |
+
"logps/rejected": -9.496674537658691,
|
| 1960 |
+
"loss": 0.1076,
|
| 1961 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 1962 |
+
"rewards/chosen": -7.032945156097412,
|
| 1963 |
+
"rewards/margins": 7.212066650390625,
|
| 1964 |
+
"rewards/rejected": -14.245012283325195,
|
| 1965 |
+
"step": 1080
|
| 1966 |
+
},
|
| 1967 |
+
{
|
| 1968 |
+
"epoch": 1.7094369236805964,
|
| 1969 |
+
"grad_norm": 4.748386859893799,
|
| 1970 |
+
"learning_rate": 8.664484900247363e-07,
|
| 1971 |
+
"logits/chosen": 11.74091911315918,
|
| 1972 |
+
"logits/rejected": 11.233245849609375,
|
| 1973 |
+
"logps/chosen": -5.176927089691162,
|
| 1974 |
+
"logps/rejected": -9.68712043762207,
|
| 1975 |
+
"loss": 0.1532,
|
| 1976 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 1977 |
+
"rewards/chosen": -7.765390872955322,
|
| 1978 |
+
"rewards/margins": 6.765290260314941,
|
| 1979 |
+
"rewards/rejected": -14.530679702758789,
|
| 1980 |
+
"step": 1090
|
| 1981 |
+
},
|
| 1982 |
+
{
|
| 1983 |
+
"epoch": 1.7251324308416716,
|
| 1984 |
+
"grad_norm": 2.932032346725464,
|
| 1985 |
+
"learning_rate": 8.271734841028553e-07,
|
| 1986 |
+
"logits/chosen": 11.53050708770752,
|
| 1987 |
+
"logits/rejected": 10.704401969909668,
|
| 1988 |
+
"logps/chosen": -5.003798484802246,
|
| 1989 |
+
"logps/rejected": -10.222925186157227,
|
| 1990 |
+
"loss": 0.1632,
|
| 1991 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 1992 |
+
"rewards/chosen": -7.505698204040527,
|
| 1993 |
+
"rewards/margins": 7.8286895751953125,
|
| 1994 |
+
"rewards/rejected": -15.334388732910156,
|
| 1995 |
+
"step": 1100
|
| 1996 |
+
},
|
| 1997 |
+
{
|
| 1998 |
+
"epoch": 1.7251324308416716,
|
| 1999 |
+
"eval_logits/chosen": 12.376419067382812,
|
| 2000 |
+
"eval_logits/rejected": 11.235088348388672,
|
| 2001 |
+
"eval_logps/chosen": -4.780559539794922,
|
| 2002 |
+
"eval_logps/rejected": -9.94357967376709,
|
| 2003 |
+
"eval_loss": 0.1563359946012497,
|
| 2004 |
+
"eval_rewards/accuracies": 0.9807692170143127,
|
| 2005 |
+
"eval_rewards/chosen": -7.170839309692383,
|
| 2006 |
+
"eval_rewards/margins": 7.744529724121094,
|
| 2007 |
+
"eval_rewards/rejected": -14.915369033813477,
|
| 2008 |
+
"eval_runtime": 14.5952,
|
| 2009 |
+
"eval_samples_per_second": 28.229,
|
| 2010 |
+
"eval_steps_per_second": 3.563,
|
| 2011 |
+
"step": 1100
|
| 2012 |
+
},
|
| 2013 |
+
{
|
| 2014 |
+
"epoch": 1.7408279380027467,
|
| 2015 |
+
"grad_norm": 3.4405391216278076,
|
| 2016 |
+
"learning_rate": 7.886322351782782e-07,
|
| 2017 |
+
"logits/chosen": 12.490205764770508,
|
| 2018 |
+
"logits/rejected": 11.409021377563477,
|
| 2019 |
+
"logps/chosen": -4.385369300842285,
|
| 2020 |
+
"logps/rejected": -9.336843490600586,
|
| 2021 |
+
"loss": 0.1579,
|
| 2022 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 2023 |
+
"rewards/chosen": -6.578053951263428,
|
| 2024 |
+
"rewards/margins": 7.427211761474609,
|
| 2025 |
+
"rewards/rejected": -14.005266189575195,
|
| 2026 |
+
"step": 1110
|
| 2027 |
+
},
|
| 2028 |
+
{
|
| 2029 |
+
"epoch": 1.7565234451638219,
|
| 2030 |
+
"grad_norm": 4.066647529602051,
|
| 2031 |
+
"learning_rate": 7.508416487165862e-07,
|
| 2032 |
+
"logits/chosen": 11.967781066894531,
|
| 2033 |
+
"logits/rejected": 11.07763671875,
|
| 2034 |
+
"logps/chosen": -4.826973915100098,
|
| 2035 |
+
"logps/rejected": -9.608440399169922,
|
| 2036 |
+
"loss": 0.1603,
|
| 2037 |
+
"rewards/accuracies": 0.9375,
|
| 2038 |
+
"rewards/chosen": -7.2404608726501465,
|
| 2039 |
+
"rewards/margins": 7.172199249267578,
|
| 2040 |
+
"rewards/rejected": -14.412660598754883,
|
| 2041 |
+
"step": 1120
|
| 2042 |
+
},
|
| 2043 |
+
{
|
| 2044 |
+
"epoch": 1.772218952324897,
|
| 2045 |
+
"grad_norm": 1.7111518383026123,
|
| 2046 |
+
"learning_rate": 7.138183009179922e-07,
|
| 2047 |
+
"logits/chosen": 12.116536140441895,
|
| 2048 |
+
"logits/rejected": 11.637681007385254,
|
| 2049 |
+
"logps/chosen": -4.51651668548584,
|
| 2050 |
+
"logps/rejected": -9.084489822387695,
|
| 2051 |
+
"loss": 0.1428,
|
| 2052 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 2053 |
+
"rewards/chosen": -6.77477502822876,
|
| 2054 |
+
"rewards/margins": 6.8519606590271,
|
| 2055 |
+
"rewards/rejected": -13.626734733581543,
|
| 2056 |
+
"step": 1130
|
| 2057 |
+
},
|
| 2058 |
+
{
|
| 2059 |
+
"epoch": 1.7879144594859722,
|
| 2060 |
+
"grad_norm": 4.302910804748535,
|
| 2061 |
+
"learning_rate": 6.775784314464717e-07,
|
| 2062 |
+
"logits/chosen": 11.709549903869629,
|
| 2063 |
+
"logits/rejected": 11.382287979125977,
|
| 2064 |
+
"logps/chosen": -5.279540061950684,
|
| 2065 |
+
"logps/rejected": -9.892831802368164,
|
| 2066 |
+
"loss": 0.163,
|
| 2067 |
+
"rewards/accuracies": 0.9375,
|
| 2068 |
+
"rewards/chosen": -7.919310569763184,
|
| 2069 |
+
"rewards/margins": 6.919938564300537,
|
| 2070 |
+
"rewards/rejected": -14.839248657226562,
|
| 2071 |
+
"step": 1140
|
| 2072 |
+
},
|
| 2073 |
+
{
|
| 2074 |
+
"epoch": 1.8036099666470473,
|
| 2075 |
+
"grad_norm": 0.8747031092643738,
|
| 2076 |
+
"learning_rate": 6.421379363065142e-07,
|
| 2077 |
+
"logits/chosen": 12.214895248413086,
|
| 2078 |
+
"logits/rejected": 11.334197998046875,
|
| 2079 |
+
"logps/chosen": -4.599240303039551,
|
| 2080 |
+
"logps/rejected": -8.940774917602539,
|
| 2081 |
+
"loss": 0.1288,
|
| 2082 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 2083 |
+
"rewards/chosen": -6.898859977722168,
|
| 2084 |
+
"rewards/margins": 6.512301445007324,
|
| 2085 |
+
"rewards/rejected": -13.411161422729492,
|
| 2086 |
+
"step": 1150
|
| 2087 |
+
},
|
| 2088 |
+
{
|
| 2089 |
+
"epoch": 1.8036099666470473,
|
| 2090 |
+
"eval_logits/chosen": 12.347259521484375,
|
| 2091 |
+
"eval_logits/rejected": 11.185004234313965,
|
| 2092 |
+
"eval_logps/chosen": -4.888545036315918,
|
| 2093 |
+
"eval_logps/rejected": -10.101993560791016,
|
| 2094 |
+
"eval_loss": 0.15526321530342102,
|
| 2095 |
+
"eval_rewards/accuracies": 0.9807692170143127,
|
| 2096 |
+
"eval_rewards/chosen": -7.332817077636719,
|
| 2097 |
+
"eval_rewards/margins": 7.8201727867126465,
|
| 2098 |
+
"eval_rewards/rejected": -15.15299129486084,
|
| 2099 |
+
"eval_runtime": 14.6049,
|
| 2100 |
+
"eval_samples_per_second": 28.21,
|
| 2101 |
+
"eval_steps_per_second": 3.56,
|
| 2102 |
+
"step": 1150
|
| 2103 |
+
},
|
| 2104 |
+
{
|
| 2105 |
+
"epoch": 1.8193054738081225,
|
| 2106 |
+
"grad_norm": 1.8233786821365356,
|
| 2107 |
+
"learning_rate": 6.075123608706093e-07,
|
| 2108 |
+
"logits/chosen": 11.725976943969727,
|
| 2109 |
+
"logits/rejected": 11.068318367004395,
|
| 2110 |
+
"logps/chosen": -5.016218185424805,
|
| 2111 |
+
"logps/rejected": -9.066827774047852,
|
| 2112 |
+
"loss": 0.1241,
|
| 2113 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 2114 |
+
"rewards/chosen": -7.524327278137207,
|
| 2115 |
+
"rewards/margins": 6.075912952423096,
|
| 2116 |
+
"rewards/rejected": -13.600239753723145,
|
| 2117 |
+
"step": 1160
|
| 2118 |
+
},
|
| 2119 |
+
{
|
| 2120 |
+
"epoch": 1.8350009809691976,
|
| 2121 |
+
"grad_norm": 1.6913511753082275,
|
| 2122 |
+
"learning_rate": 5.737168930605272e-07,
|
| 2123 |
+
"logits/chosen": 11.813165664672852,
|
| 2124 |
+
"logits/rejected": 10.936023712158203,
|
| 2125 |
+
"logps/chosen": -5.16327428817749,
|
| 2126 |
+
"logps/rejected": -9.918088912963867,
|
| 2127 |
+
"loss": 0.1337,
|
| 2128 |
+
"rewards/accuracies": 0.9750000238418579,
|
| 2129 |
+
"rewards/chosen": -7.744911193847656,
|
| 2130 |
+
"rewards/margins": 7.132222652435303,
|
| 2131 |
+
"rewards/rejected": -14.877134323120117,
|
| 2132 |
+
"step": 1170
|
| 2133 |
+
},
|
| 2134 |
+
{
|
| 2135 |
+
"epoch": 1.8506964881302728,
|
| 2136 |
+
"grad_norm": 4.435967445373535,
|
| 2137 |
+
"learning_rate": 5.407663566854008e-07,
|
| 2138 |
+
"logits/chosen": 11.606057167053223,
|
| 2139 |
+
"logits/rejected": 10.97921085357666,
|
| 2140 |
+
"logps/chosen": -5.703640937805176,
|
| 2141 |
+
"logps/rejected": -10.870219230651855,
|
| 2142 |
+
"loss": 0.1245,
|
| 2143 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 2144 |
+
"rewards/chosen": -8.555461883544922,
|
| 2145 |
+
"rewards/margins": 7.749865531921387,
|
| 2146 |
+
"rewards/rejected": -16.305328369140625,
|
| 2147 |
+
"step": 1180
|
| 2148 |
+
},
|
| 2149 |
+
{
|
| 2150 |
+
"epoch": 1.8663919952913477,
|
| 2151 |
+
"grad_norm": 2.9056365489959717,
|
| 2152 |
+
"learning_rate": 5.086752049395094e-07,
|
| 2153 |
+
"logits/chosen": 11.697361946105957,
|
| 2154 |
+
"logits/rejected": 10.928617477416992,
|
| 2155 |
+
"logps/chosen": -5.48267936706543,
|
| 2156 |
+
"logps/rejected": -10.689864158630371,
|
| 2157 |
+
"loss": 0.1251,
|
| 2158 |
+
"rewards/accuracies": 0.987500011920929,
|
| 2159 |
+
"rewards/chosen": -8.224019050598145,
|
| 2160 |
+
"rewards/margins": 7.810777187347412,
|
| 2161 |
+
"rewards/rejected": -16.0347957611084,
|
| 2162 |
+
"step": 1190
|
| 2163 |
+
},
|
| 2164 |
+
{
|
| 2165 |
+
"epoch": 1.882087502452423,
|
| 2166 |
+
"grad_norm": 2.538482904434204,
|
| 2167 |
+
"learning_rate": 4.774575140626317e-07,
|
| 2168 |
+
"logits/chosen": 11.8221435546875,
|
| 2169 |
+
"logits/rejected": 11.30038070678711,
|
| 2170 |
+
"logps/chosen": -5.215182781219482,
|
| 2171 |
+
"logps/rejected": -9.782806396484375,
|
| 2172 |
+
"loss": 0.1543,
|
| 2173 |
+
"rewards/accuracies": 0.9624999761581421,
|
| 2174 |
+
"rewards/chosen": -7.822774410247803,
|
| 2175 |
+
"rewards/margins": 6.851436614990234,
|
| 2176 |
+
"rewards/rejected": -14.674212455749512,
|
| 2177 |
+
"step": 1200
|
| 2178 |
+
},
|
| 2179 |
+
{
|
| 2180 |
+
"epoch": 1.882087502452423,
|
| 2181 |
+
"eval_logits/chosen": 12.300190925598145,
|
| 2182 |
+
"eval_logits/rejected": 11.128490447998047,
|
| 2183 |
+
"eval_logps/chosen": -4.917614936828613,
|
| 2184 |
+
"eval_logps/rejected": -10.187230110168457,
|
| 2185 |
+
"eval_loss": 0.1518644094467163,
|
| 2186 |
+
"eval_rewards/accuracies": 0.9807692170143127,
|
| 2187 |
+
"eval_rewards/chosen": -7.37642240524292,
|
| 2188 |
+
"eval_rewards/margins": 7.904422283172607,
|
| 2189 |
+
"eval_rewards/rejected": -15.280843734741211,
|
| 2190 |
+
"eval_runtime": 14.6081,
|
| 2191 |
+
"eval_samples_per_second": 28.203,
|
| 2192 |
+
"eval_steps_per_second": 3.56,
|
| 2193 |
+
"step": 1200
|
| 2194 |
+
}
|
| 2195 |
+
],
|
| 2196 |
+
"logging_steps": 10,
|
| 2197 |
+
"max_steps": 1500,
|
| 2198 |
+
"num_input_tokens_seen": 0,
|
| 2199 |
+
"num_train_epochs": 3,
|
| 2200 |
+
"save_steps": 50,
|
| 2201 |
+
"stateful_callbacks": {
|
| 2202 |
+
"TrainerControl": {
|
| 2203 |
+
"args": {
|
| 2204 |
+
"should_epoch_stop": false,
|
| 2205 |
+
"should_evaluate": false,
|
| 2206 |
+
"should_log": false,
|
| 2207 |
+
"should_save": true,
|
| 2208 |
+
"should_training_stop": false
|
| 2209 |
+
},
|
| 2210 |
+
"attributes": {}
|
| 2211 |
+
}
|
| 2212 |
+
},
|
| 2213 |
+
"total_flos": 2.8694519974492897e+18,
|
| 2214 |
+
"train_batch_size": 1,
|
| 2215 |
+
"trial_name": null,
|
| 2216 |
+
"trial_params": null
|
| 2217 |
+
}
|
checkpoint-1200/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5163ac022a6b5f37a143d69c025be3610fb33a65145b769bdd1d9b31cc494442
|
| 3 |
+
size 7224
|
checkpoint-1200/zero_to_fp32.py
ADDED
|
@@ -0,0 +1,674 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
#!/usr/bin/env python
|
| 2 |
+
|
| 3 |
+
# Copyright (c) Microsoft Corporation.
|
| 4 |
+
# SPDX-License-Identifier: Apache-2.0
|
| 5 |
+
|
| 6 |
+
# DeepSpeed Team
|
| 7 |
+
|
| 8 |
+
# This script extracts fp32 consolidated weights from a zero 1, 2 and 3 DeepSpeed checkpoints. It gets
|
| 9 |
+
# copied into the top level checkpoint dir, so the user can easily do the conversion at any point in
|
| 10 |
+
# the future. Once extracted, the weights don't require DeepSpeed and can be used in any
|
| 11 |
+
# application.
|
| 12 |
+
#
|
| 13 |
+
# example:
|
| 14 |
+
# python zero_to_fp32.py . output_dir/
|
| 15 |
+
# or
|
| 16 |
+
# python zero_to_fp32.py . output_dir/ --safe_serialization
|
| 17 |
+
|
| 18 |
+
import argparse
|
| 19 |
+
import torch
|
| 20 |
+
import glob
|
| 21 |
+
import math
|
| 22 |
+
import os
|
| 23 |
+
import re
|
| 24 |
+
import json
|
| 25 |
+
from tqdm import tqdm
|
| 26 |
+
from collections import OrderedDict
|
| 27 |
+
from dataclasses import dataclass
|
| 28 |
+
|
| 29 |
+
# while this script doesn't use deepspeed to recover data, since the checkpoints are pickled with
|
| 30 |
+
# DeepSpeed data structures it has to be available in the current python environment.
|
| 31 |
+
from deepspeed.utils import logger
|
| 32 |
+
from deepspeed.checkpoint.constants import (DS_VERSION, OPTIMIZER_STATE_DICT, SINGLE_PARTITION_OF_FP32_GROUPS,
|
| 33 |
+
FP32_FLAT_GROUPS, ZERO_STAGE, PARTITION_COUNT, PARAM_SHAPES, BUFFER_NAMES,
|
| 34 |
+
FROZEN_PARAM_SHAPES, FROZEN_PARAM_FRAGMENTS)
|
| 35 |
+
|
| 36 |
+
|
| 37 |
+
@dataclass
|
| 38 |
+
class zero_model_state:
|
| 39 |
+
buffers: dict()
|
| 40 |
+
param_shapes: dict()
|
| 41 |
+
shared_params: list
|
| 42 |
+
ds_version: int
|
| 43 |
+
frozen_param_shapes: dict()
|
| 44 |
+
frozen_param_fragments: dict()
|
| 45 |
+
|
| 46 |
+
|
| 47 |
+
debug = 0
|
| 48 |
+
|
| 49 |
+
# load to cpu
|
| 50 |
+
device = torch.device('cpu')
|
| 51 |
+
|
| 52 |
+
|
| 53 |
+
def atoi(text):
|
| 54 |
+
return int(text) if text.isdigit() else text
|
| 55 |
+
|
| 56 |
+
|
| 57 |
+
def natural_keys(text):
|
| 58 |
+
'''
|
| 59 |
+
alist.sort(key=natural_keys) sorts in human order
|
| 60 |
+
http://nedbatchelder.com/blog/200712/human_sorting.html
|
| 61 |
+
(See Toothy's implementation in the comments)
|
| 62 |
+
'''
|
| 63 |
+
return [atoi(c) for c in re.split(r'(\d+)', text)]
|
| 64 |
+
|
| 65 |
+
|
| 66 |
+
def get_model_state_file(checkpoint_dir, zero_stage):
|
| 67 |
+
if not os.path.isdir(checkpoint_dir):
|
| 68 |
+
raise FileNotFoundError(f"Directory '{checkpoint_dir}' doesn't exist")
|
| 69 |
+
|
| 70 |
+
# there should be only one file
|
| 71 |
+
if zero_stage <= 2:
|
| 72 |
+
file = os.path.join(checkpoint_dir, "mp_rank_00_model_states.pt")
|
| 73 |
+
elif zero_stage == 3:
|
| 74 |
+
file = os.path.join(checkpoint_dir, "zero_pp_rank_0_mp_rank_00_model_states.pt")
|
| 75 |
+
|
| 76 |
+
if not os.path.exists(file):
|
| 77 |
+
raise FileNotFoundError(f"can't find model states file at '{file}'")
|
| 78 |
+
|
| 79 |
+
return file
|
| 80 |
+
|
| 81 |
+
|
| 82 |
+
def get_checkpoint_files(checkpoint_dir, glob_pattern):
|
| 83 |
+
# XXX: need to test that this simple glob rule works for multi-node setup too
|
| 84 |
+
ckpt_files = sorted(glob.glob(os.path.join(checkpoint_dir, glob_pattern)), key=natural_keys)
|
| 85 |
+
|
| 86 |
+
if len(ckpt_files) == 0:
|
| 87 |
+
raise FileNotFoundError(f"can't find {glob_pattern} files in directory '{checkpoint_dir}'")
|
| 88 |
+
|
| 89 |
+
return ckpt_files
|
| 90 |
+
|
| 91 |
+
|
| 92 |
+
def get_optim_files(checkpoint_dir):
|
| 93 |
+
return get_checkpoint_files(checkpoint_dir, "*_optim_states.pt")
|
| 94 |
+
|
| 95 |
+
|
| 96 |
+
def get_model_state_files(checkpoint_dir):
|
| 97 |
+
return get_checkpoint_files(checkpoint_dir, "*_model_states.pt")
|
| 98 |
+
|
| 99 |
+
|
| 100 |
+
def parse_model_states(files):
|
| 101 |
+
zero_model_states = []
|
| 102 |
+
for file in files:
|
| 103 |
+
state_dict = torch.load(file, map_location=device)
|
| 104 |
+
|
| 105 |
+
if BUFFER_NAMES not in state_dict:
|
| 106 |
+
raise ValueError(f"{file} is not a model state checkpoint")
|
| 107 |
+
buffer_names = state_dict[BUFFER_NAMES]
|
| 108 |
+
if debug:
|
| 109 |
+
print("Found buffers:", buffer_names)
|
| 110 |
+
|
| 111 |
+
# recover just the buffers while restoring them to fp32 if they were saved in fp16
|
| 112 |
+
buffers = {k: v.float() for k, v in state_dict["module"].items() if k in buffer_names}
|
| 113 |
+
param_shapes = state_dict[PARAM_SHAPES]
|
| 114 |
+
|
| 115 |
+
# collect parameters that are included in param_shapes
|
| 116 |
+
param_names = []
|
| 117 |
+
for s in param_shapes:
|
| 118 |
+
for name in s.keys():
|
| 119 |
+
param_names.append(name)
|
| 120 |
+
|
| 121 |
+
# update with frozen parameters
|
| 122 |
+
frozen_param_shapes = state_dict.get(FROZEN_PARAM_SHAPES, None)
|
| 123 |
+
if frozen_param_shapes is not None:
|
| 124 |
+
if debug:
|
| 125 |
+
print(f"Found frozen_param_shapes: {frozen_param_shapes}")
|
| 126 |
+
param_names += list(frozen_param_shapes.keys())
|
| 127 |
+
|
| 128 |
+
# handle shared params
|
| 129 |
+
shared_params = [[k, v] for k, v in state_dict["shared_params"].items()]
|
| 130 |
+
|
| 131 |
+
ds_version = state_dict.get(DS_VERSION, None)
|
| 132 |
+
|
| 133 |
+
frozen_param_fragments = state_dict.get(FROZEN_PARAM_FRAGMENTS, None)
|
| 134 |
+
|
| 135 |
+
z_model_state = zero_model_state(buffers=buffers,
|
| 136 |
+
param_shapes=param_shapes,
|
| 137 |
+
shared_params=shared_params,
|
| 138 |
+
ds_version=ds_version,
|
| 139 |
+
frozen_param_shapes=frozen_param_shapes,
|
| 140 |
+
frozen_param_fragments=frozen_param_fragments)
|
| 141 |
+
zero_model_states.append(z_model_state)
|
| 142 |
+
|
| 143 |
+
return zero_model_states
|
| 144 |
+
|
| 145 |
+
|
| 146 |
+
def parse_optim_states(files, ds_checkpoint_dir):
|
| 147 |
+
total_files = len(files)
|
| 148 |
+
state_dicts = []
|
| 149 |
+
for f in files:
|
| 150 |
+
state_dict = torch.load(f, map_location=device)
|
| 151 |
+
# immediately discard the potentially huge 2 optimizer states as we only care for fp32 master weights
|
| 152 |
+
# and also handle the case where it was already removed by another helper script
|
| 153 |
+
state_dict["optimizer_state_dict"].pop("optimizer_state_dict", None)
|
| 154 |
+
state_dicts.append(state_dict)
|
| 155 |
+
|
| 156 |
+
if not ZERO_STAGE in state_dicts[0][OPTIMIZER_STATE_DICT]:
|
| 157 |
+
raise ValueError(f"{files[0]} is not a zero checkpoint")
|
| 158 |
+
zero_stage = state_dicts[0][OPTIMIZER_STATE_DICT][ZERO_STAGE]
|
| 159 |
+
world_size = state_dicts[0][OPTIMIZER_STATE_DICT][PARTITION_COUNT]
|
| 160 |
+
|
| 161 |
+
# For ZeRO-2 each param group can have different partition_count as data parallelism for expert
|
| 162 |
+
# parameters can be different from data parallelism for non-expert parameters. So we can just
|
| 163 |
+
# use the max of the partition_count to get the dp world_size.
|
| 164 |
+
|
| 165 |
+
if type(world_size) is list:
|
| 166 |
+
world_size = max(world_size)
|
| 167 |
+
|
| 168 |
+
if world_size != total_files:
|
| 169 |
+
raise ValueError(
|
| 170 |
+
f"Expected {world_size} of '*_optim_states.pt' under '{ds_checkpoint_dir}' but found {total_files} files. "
|
| 171 |
+
"Possibly due to an overwrite of an old checkpoint, or a checkpoint didn't get saved by one or more processes."
|
| 172 |
+
)
|
| 173 |
+
|
| 174 |
+
# the groups are named differently in each stage
|
| 175 |
+
if zero_stage <= 2:
|
| 176 |
+
fp32_groups_key = SINGLE_PARTITION_OF_FP32_GROUPS
|
| 177 |
+
elif zero_stage == 3:
|
| 178 |
+
fp32_groups_key = FP32_FLAT_GROUPS
|
| 179 |
+
else:
|
| 180 |
+
raise ValueError(f"unknown zero stage {zero_stage}")
|
| 181 |
+
|
| 182 |
+
if zero_stage <= 2:
|
| 183 |
+
fp32_flat_groups = [state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key] for i in range(len(state_dicts))]
|
| 184 |
+
elif zero_stage == 3:
|
| 185 |
+
# if there is more than one param group, there will be multiple flattened tensors - one
|
| 186 |
+
# flattened tensor per group - for simplicity merge them into a single tensor
|
| 187 |
+
#
|
| 188 |
+
# XXX: could make the script more memory efficient for when there are multiple groups - it
|
| 189 |
+
# will require matching the sub-lists of param_shapes for each param group flattened tensor
|
| 190 |
+
|
| 191 |
+
fp32_flat_groups = [
|
| 192 |
+
torch.cat(state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key], 0) for i in range(len(state_dicts))
|
| 193 |
+
]
|
| 194 |
+
|
| 195 |
+
return zero_stage, world_size, fp32_flat_groups
|
| 196 |
+
|
| 197 |
+
|
| 198 |
+
def _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters):
|
| 199 |
+
"""
|
| 200 |
+
Returns fp32 state_dict reconstructed from ds checkpoint
|
| 201 |
+
|
| 202 |
+
Args:
|
| 203 |
+
- ``ds_checkpoint_dir``: path to the deepspeed checkpoint folder (where the optimizer files are)
|
| 204 |
+
|
| 205 |
+
"""
|
| 206 |
+
print(f"Processing zero checkpoint '{ds_checkpoint_dir}'")
|
| 207 |
+
|
| 208 |
+
optim_files = get_optim_files(ds_checkpoint_dir)
|
| 209 |
+
zero_stage, world_size, fp32_flat_groups = parse_optim_states(optim_files, ds_checkpoint_dir)
|
| 210 |
+
print(f"Detected checkpoint of type zero stage {zero_stage}, world_size: {world_size}")
|
| 211 |
+
|
| 212 |
+
model_files = get_model_state_files(ds_checkpoint_dir)
|
| 213 |
+
|
| 214 |
+
zero_model_states = parse_model_states(model_files)
|
| 215 |
+
print(f'Parsing checkpoint created by deepspeed=={zero_model_states[0].ds_version}')
|
| 216 |
+
|
| 217 |
+
if zero_stage <= 2:
|
| 218 |
+
return _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states,
|
| 219 |
+
exclude_frozen_parameters)
|
| 220 |
+
elif zero_stage == 3:
|
| 221 |
+
return _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states,
|
| 222 |
+
exclude_frozen_parameters)
|
| 223 |
+
|
| 224 |
+
|
| 225 |
+
def _zero2_merge_frozen_params(state_dict, zero_model_states):
|
| 226 |
+
if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
|
| 227 |
+
return
|
| 228 |
+
|
| 229 |
+
frozen_param_shapes = zero_model_states[0].frozen_param_shapes
|
| 230 |
+
frozen_param_fragments = zero_model_states[0].frozen_param_fragments
|
| 231 |
+
|
| 232 |
+
if debug:
|
| 233 |
+
num_elem = sum(s.numel() for s in frozen_param_shapes.values())
|
| 234 |
+
print(f'rank 0: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
|
| 235 |
+
|
| 236 |
+
wanted_params = len(frozen_param_shapes)
|
| 237 |
+
wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
|
| 238 |
+
avail_numel = sum([p.numel() for p in frozen_param_fragments.values()])
|
| 239 |
+
print(f'Frozen params: Have {avail_numel} numels to process.')
|
| 240 |
+
print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
|
| 241 |
+
|
| 242 |
+
total_params = 0
|
| 243 |
+
total_numel = 0
|
| 244 |
+
for name, shape in frozen_param_shapes.items():
|
| 245 |
+
total_params += 1
|
| 246 |
+
unpartitioned_numel = shape.numel()
|
| 247 |
+
total_numel += unpartitioned_numel
|
| 248 |
+
|
| 249 |
+
state_dict[name] = frozen_param_fragments[name]
|
| 250 |
+
|
| 251 |
+
if debug:
|
| 252 |
+
print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
|
| 253 |
+
|
| 254 |
+
print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
|
| 255 |
+
|
| 256 |
+
|
| 257 |
+
def _has_callable(obj, fn):
|
| 258 |
+
attr = getattr(obj, fn, None)
|
| 259 |
+
return callable(attr)
|
| 260 |
+
|
| 261 |
+
|
| 262 |
+
def _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
|
| 263 |
+
param_shapes = zero_model_states[0].param_shapes
|
| 264 |
+
|
| 265 |
+
# Reconstruction protocol:
|
| 266 |
+
#
|
| 267 |
+
# XXX: document this
|
| 268 |
+
|
| 269 |
+
if debug:
|
| 270 |
+
for i in range(world_size):
|
| 271 |
+
for j in range(len(fp32_flat_groups[0])):
|
| 272 |
+
print(f"{FP32_FLAT_GROUPS}[{i}][{j}].shape={fp32_flat_groups[i][j].shape}")
|
| 273 |
+
|
| 274 |
+
# XXX: memory usage doubles here (zero2)
|
| 275 |
+
num_param_groups = len(fp32_flat_groups[0])
|
| 276 |
+
merged_single_partition_of_fp32_groups = []
|
| 277 |
+
for i in range(num_param_groups):
|
| 278 |
+
merged_partitions = [sd[i] for sd in fp32_flat_groups]
|
| 279 |
+
full_single_fp32_vector = torch.cat(merged_partitions, 0)
|
| 280 |
+
merged_single_partition_of_fp32_groups.append(full_single_fp32_vector)
|
| 281 |
+
avail_numel = sum(
|
| 282 |
+
[full_single_fp32_vector.numel() for full_single_fp32_vector in merged_single_partition_of_fp32_groups])
|
| 283 |
+
|
| 284 |
+
if debug:
|
| 285 |
+
wanted_params = sum([len(shapes) for shapes in param_shapes])
|
| 286 |
+
wanted_numel = sum([sum(shape.numel() for shape in shapes.values()) for shapes in param_shapes])
|
| 287 |
+
# not asserting if there is a mismatch due to possible padding
|
| 288 |
+
print(f"Have {avail_numel} numels to process.")
|
| 289 |
+
print(f"Need {wanted_numel} numels in {wanted_params} params.")
|
| 290 |
+
|
| 291 |
+
# params
|
| 292 |
+
# XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
|
| 293 |
+
# out-of-core computing solution
|
| 294 |
+
total_numel = 0
|
| 295 |
+
total_params = 0
|
| 296 |
+
for shapes, full_single_fp32_vector in zip(param_shapes, merged_single_partition_of_fp32_groups):
|
| 297 |
+
offset = 0
|
| 298 |
+
avail_numel = full_single_fp32_vector.numel()
|
| 299 |
+
for name, shape in shapes.items():
|
| 300 |
+
|
| 301 |
+
unpartitioned_numel = shape.numel() if _has_callable(shape, 'numel') else math.prod(shape)
|
| 302 |
+
total_numel += unpartitioned_numel
|
| 303 |
+
total_params += 1
|
| 304 |
+
|
| 305 |
+
if debug:
|
| 306 |
+
print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
|
| 307 |
+
state_dict[name] = full_single_fp32_vector.narrow(0, offset, unpartitioned_numel).view(shape)
|
| 308 |
+
offset += unpartitioned_numel
|
| 309 |
+
|
| 310 |
+
# Z2 started to align to 2*world_size to improve nccl performance. Therefore both offset and
|
| 311 |
+
# avail_numel can differ by anywhere between 0..2*world_size. Due to two unrelated complex
|
| 312 |
+
# paddings performed in the code it's almost impossible to predict the exact numbers w/o the
|
| 313 |
+
# live optimizer object, so we are checking that the numbers are within the right range
|
| 314 |
+
align_to = 2 * world_size
|
| 315 |
+
|
| 316 |
+
def zero2_align(x):
|
| 317 |
+
return align_to * math.ceil(x / align_to)
|
| 318 |
+
|
| 319 |
+
if debug:
|
| 320 |
+
print(f"original offset={offset}, avail_numel={avail_numel}")
|
| 321 |
+
|
| 322 |
+
offset = zero2_align(offset)
|
| 323 |
+
avail_numel = zero2_align(avail_numel)
|
| 324 |
+
|
| 325 |
+
if debug:
|
| 326 |
+
print(f"aligned offset={offset}, avail_numel={avail_numel}")
|
| 327 |
+
|
| 328 |
+
# Sanity check
|
| 329 |
+
if offset != avail_numel:
|
| 330 |
+
raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
|
| 331 |
+
|
| 332 |
+
print(f"Reconstructed fp32 state dict with {total_params} params {total_numel} elements")
|
| 333 |
+
|
| 334 |
+
|
| 335 |
+
def _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states,
|
| 336 |
+
exclude_frozen_parameters):
|
| 337 |
+
state_dict = OrderedDict()
|
| 338 |
+
|
| 339 |
+
# buffers
|
| 340 |
+
buffers = zero_model_states[0].buffers
|
| 341 |
+
state_dict.update(buffers)
|
| 342 |
+
if debug:
|
| 343 |
+
print(f"added {len(buffers)} buffers")
|
| 344 |
+
|
| 345 |
+
if not exclude_frozen_parameters:
|
| 346 |
+
_zero2_merge_frozen_params(state_dict, zero_model_states)
|
| 347 |
+
|
| 348 |
+
_zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
|
| 349 |
+
|
| 350 |
+
# recover shared parameters
|
| 351 |
+
for pair in zero_model_states[0].shared_params:
|
| 352 |
+
if pair[1] in state_dict:
|
| 353 |
+
state_dict[pair[0]] = state_dict[pair[1]]
|
| 354 |
+
|
| 355 |
+
return state_dict
|
| 356 |
+
|
| 357 |
+
|
| 358 |
+
def zero3_partitioned_param_info(unpartitioned_numel, world_size):
|
| 359 |
+
remainder = unpartitioned_numel % world_size
|
| 360 |
+
padding_numel = (world_size - remainder) if remainder else 0
|
| 361 |
+
partitioned_numel = math.ceil(unpartitioned_numel / world_size)
|
| 362 |
+
return partitioned_numel, padding_numel
|
| 363 |
+
|
| 364 |
+
|
| 365 |
+
def _zero3_merge_frozen_params(state_dict, world_size, zero_model_states):
|
| 366 |
+
if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
|
| 367 |
+
return
|
| 368 |
+
|
| 369 |
+
if debug:
|
| 370 |
+
for i in range(world_size):
|
| 371 |
+
num_elem = sum(s.numel() for s in zero_model_states[i].frozen_param_fragments.values())
|
| 372 |
+
print(f'rank {i}: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
|
| 373 |
+
|
| 374 |
+
frozen_param_shapes = zero_model_states[0].frozen_param_shapes
|
| 375 |
+
wanted_params = len(frozen_param_shapes)
|
| 376 |
+
wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
|
| 377 |
+
avail_numel = sum([p.numel() for p in zero_model_states[0].frozen_param_fragments.values()]) * world_size
|
| 378 |
+
print(f'Frozen params: Have {avail_numel} numels to process.')
|
| 379 |
+
print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
|
| 380 |
+
|
| 381 |
+
total_params = 0
|
| 382 |
+
total_numel = 0
|
| 383 |
+
for name, shape in zero_model_states[0].frozen_param_shapes.items():
|
| 384 |
+
total_params += 1
|
| 385 |
+
unpartitioned_numel = shape.numel()
|
| 386 |
+
total_numel += unpartitioned_numel
|
| 387 |
+
|
| 388 |
+
param_frags = tuple(model_state.frozen_param_fragments[name] for model_state in zero_model_states)
|
| 389 |
+
state_dict[name] = torch.cat(param_frags, 0).narrow(0, 0, unpartitioned_numel).view(shape)
|
| 390 |
+
|
| 391 |
+
partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
|
| 392 |
+
|
| 393 |
+
if debug:
|
| 394 |
+
print(
|
| 395 |
+
f"Frozen params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
|
| 396 |
+
)
|
| 397 |
+
|
| 398 |
+
print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
|
| 399 |
+
|
| 400 |
+
|
| 401 |
+
def _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
|
| 402 |
+
param_shapes = zero_model_states[0].param_shapes
|
| 403 |
+
avail_numel = fp32_flat_groups[0].numel() * world_size
|
| 404 |
+
# Reconstruction protocol: For zero3 we need to zip the partitions together at boundary of each
|
| 405 |
+
# param, re-consolidating each param, while dealing with padding if any
|
| 406 |
+
|
| 407 |
+
# merge list of dicts, preserving order
|
| 408 |
+
param_shapes = {k: v for d in param_shapes for k, v in d.items()}
|
| 409 |
+
|
| 410 |
+
if debug:
|
| 411 |
+
for i in range(world_size):
|
| 412 |
+
print(f"{FP32_FLAT_GROUPS}[{i}].shape={fp32_flat_groups[i].shape}")
|
| 413 |
+
|
| 414 |
+
wanted_params = len(param_shapes)
|
| 415 |
+
wanted_numel = sum(shape.numel() for shape in param_shapes.values())
|
| 416 |
+
# not asserting if there is a mismatch due to possible padding
|
| 417 |
+
avail_numel = fp32_flat_groups[0].numel() * world_size
|
| 418 |
+
print(f"Trainable params: Have {avail_numel} numels to process.")
|
| 419 |
+
print(f"Trainable params: Need {wanted_numel} numels in {wanted_params} params.")
|
| 420 |
+
|
| 421 |
+
# params
|
| 422 |
+
# XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
|
| 423 |
+
# out-of-core computing solution
|
| 424 |
+
offset = 0
|
| 425 |
+
total_numel = 0
|
| 426 |
+
total_params = 0
|
| 427 |
+
for name, shape in tqdm(param_shapes.items(), desc='Gathering Sharded Weights'):
|
| 428 |
+
unpartitioned_numel = shape.numel()
|
| 429 |
+
total_numel += unpartitioned_numel
|
| 430 |
+
total_params += 1
|
| 431 |
+
partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
|
| 432 |
+
|
| 433 |
+
if debug:
|
| 434 |
+
print(
|
| 435 |
+
f"Trainable params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
|
| 436 |
+
)
|
| 437 |
+
|
| 438 |
+
# XXX: memory usage doubles here
|
| 439 |
+
state_dict[name] = torch.cat(
|
| 440 |
+
tuple(fp32_flat_groups[i].narrow(0, offset, partitioned_numel) for i in range(world_size)),
|
| 441 |
+
0).narrow(0, 0, unpartitioned_numel).view(shape)
|
| 442 |
+
offset += partitioned_numel
|
| 443 |
+
|
| 444 |
+
offset *= world_size
|
| 445 |
+
|
| 446 |
+
# Sanity check
|
| 447 |
+
if offset != avail_numel:
|
| 448 |
+
raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
|
| 449 |
+
|
| 450 |
+
print(f"Reconstructed Trainable fp32 state dict with {total_params} params {total_numel} elements")
|
| 451 |
+
|
| 452 |
+
|
| 453 |
+
def _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states,
|
| 454 |
+
exclude_frozen_parameters):
|
| 455 |
+
state_dict = OrderedDict()
|
| 456 |
+
|
| 457 |
+
# buffers
|
| 458 |
+
buffers = zero_model_states[0].buffers
|
| 459 |
+
state_dict.update(buffers)
|
| 460 |
+
if debug:
|
| 461 |
+
print(f"added {len(buffers)} buffers")
|
| 462 |
+
|
| 463 |
+
if not exclude_frozen_parameters:
|
| 464 |
+
_zero3_merge_frozen_params(state_dict, world_size, zero_model_states)
|
| 465 |
+
|
| 466 |
+
_zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
|
| 467 |
+
|
| 468 |
+
# recover shared parameters
|
| 469 |
+
for pair in zero_model_states[0].shared_params:
|
| 470 |
+
if pair[1] in state_dict:
|
| 471 |
+
state_dict[pair[0]] = state_dict[pair[1]]
|
| 472 |
+
|
| 473 |
+
return state_dict
|
| 474 |
+
|
| 475 |
+
|
| 476 |
+
def get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag=None, exclude_frozen_parameters=False):
|
| 477 |
+
"""
|
| 478 |
+
Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated state_dict that can be loaded with
|
| 479 |
+
``load_state_dict()`` and used for training without DeepSpeed or shared with others, for example
|
| 480 |
+
via a model hub.
|
| 481 |
+
|
| 482 |
+
Args:
|
| 483 |
+
- ``checkpoint_dir``: path to the desired checkpoint folder
|
| 484 |
+
- ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in 'latest' file. e.g., ``global_step14``
|
| 485 |
+
- ``exclude_frozen_parameters``: exclude frozen parameters
|
| 486 |
+
|
| 487 |
+
Returns:
|
| 488 |
+
- pytorch ``state_dict``
|
| 489 |
+
|
| 490 |
+
Note: this approach may not work if your application doesn't have sufficient free CPU memory and
|
| 491 |
+
you may need to use the offline approach using the ``zero_to_fp32.py`` script that is saved with
|
| 492 |
+
the checkpoint.
|
| 493 |
+
|
| 494 |
+
A typical usage might be ::
|
| 495 |
+
|
| 496 |
+
from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint
|
| 497 |
+
# do the training and checkpoint saving
|
| 498 |
+
state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir) # already on cpu
|
| 499 |
+
model = model.cpu() # move to cpu
|
| 500 |
+
model.load_state_dict(state_dict)
|
| 501 |
+
# submit to model hub or save the model to share with others
|
| 502 |
+
|
| 503 |
+
In this example the ``model`` will no longer be usable in the deepspeed context of the same
|
| 504 |
+
application. i.e. you will need to re-initialize the deepspeed engine, since
|
| 505 |
+
``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
|
| 506 |
+
|
| 507 |
+
If you want it all done for you, use ``load_state_dict_from_zero_checkpoint`` instead.
|
| 508 |
+
|
| 509 |
+
"""
|
| 510 |
+
if tag is None:
|
| 511 |
+
latest_path = os.path.join(checkpoint_dir, 'latest')
|
| 512 |
+
if os.path.isfile(latest_path):
|
| 513 |
+
with open(latest_path, 'r') as fd:
|
| 514 |
+
tag = fd.read().strip()
|
| 515 |
+
else:
|
| 516 |
+
raise ValueError(f"Unable to find 'latest' file at {latest_path}")
|
| 517 |
+
|
| 518 |
+
ds_checkpoint_dir = os.path.join(checkpoint_dir, tag)
|
| 519 |
+
|
| 520 |
+
if not os.path.isdir(ds_checkpoint_dir):
|
| 521 |
+
raise FileNotFoundError(f"Directory '{ds_checkpoint_dir}' doesn't exist")
|
| 522 |
+
|
| 523 |
+
return _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters)
|
| 524 |
+
|
| 525 |
+
|
| 526 |
+
def convert_zero_checkpoint_to_fp32_state_dict(checkpoint_dir,
|
| 527 |
+
output_dir,
|
| 528 |
+
max_shard_size="5GB",
|
| 529 |
+
safe_serialization=False,
|
| 530 |
+
tag=None,
|
| 531 |
+
exclude_frozen_parameters=False):
|
| 532 |
+
"""
|
| 533 |
+
Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` file that can be
|
| 534 |
+
loaded with ``torch.load(file)`` + ``load_state_dict()`` and used for training without DeepSpeed.
|
| 535 |
+
|
| 536 |
+
Args:
|
| 537 |
+
- ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
|
| 538 |
+
- ``output_dir``: directory to the pytorch fp32 state_dict output files
|
| 539 |
+
- ``max_shard_size``: the maximum size for a checkpoint before being sharded, default value is 5GB
|
| 540 |
+
- ``safe_serialization``: whether to save the model using `safetensors` or the traditional PyTorch way (that uses `pickle`).
|
| 541 |
+
- ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
|
| 542 |
+
- ``exclude_frozen_parameters``: exclude frozen parameters
|
| 543 |
+
"""
|
| 544 |
+
# Dependency pre-check
|
| 545 |
+
if safe_serialization:
|
| 546 |
+
try:
|
| 547 |
+
from safetensors.torch import save_file
|
| 548 |
+
except ImportError:
|
| 549 |
+
print('If you want to use `safe_serialization`, please `pip install safetensors`')
|
| 550 |
+
raise
|
| 551 |
+
if max_shard_size is not None:
|
| 552 |
+
try:
|
| 553 |
+
from huggingface_hub import split_torch_state_dict_into_shards
|
| 554 |
+
except ImportError:
|
| 555 |
+
print('If you want to use `max_shard_size`, please `pip install huggingface_hub`')
|
| 556 |
+
raise
|
| 557 |
+
|
| 558 |
+
# Convert zero checkpoint to state_dict
|
| 559 |
+
state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag, exclude_frozen_parameters)
|
| 560 |
+
|
| 561 |
+
# Shard the model if it is too big.
|
| 562 |
+
weights_name = "model.safetensors" if safe_serialization else "pytorch_model.bin"
|
| 563 |
+
if max_shard_size is not None:
|
| 564 |
+
filename_pattern = weights_name.replace(".bin", "{suffix}.bin").replace(".safetensors", "{suffix}.safetensors")
|
| 565 |
+
state_dict_split = split_torch_state_dict_into_shards(state_dict,
|
| 566 |
+
filename_pattern=filename_pattern,
|
| 567 |
+
max_shard_size=max_shard_size)
|
| 568 |
+
else:
|
| 569 |
+
from collections import namedtuple
|
| 570 |
+
StateDictSplit = namedtuple("StateDictSplit", ["is_sharded", "filename_to_tensors"])
|
| 571 |
+
state_dict_split = StateDictSplit(is_sharded=False,
|
| 572 |
+
filename_to_tensors={weights_name: list(state_dict.keys())})
|
| 573 |
+
|
| 574 |
+
# Save the model
|
| 575 |
+
filename_to_tensors = state_dict_split.filename_to_tensors.items()
|
| 576 |
+
for shard_file, tensors in tqdm(filename_to_tensors, desc="Saving checkpoint shards"):
|
| 577 |
+
shard = {tensor: state_dict[tensor].contiguous() for tensor in tensors}
|
| 578 |
+
output_path = os.path.join(output_dir, shard_file)
|
| 579 |
+
if safe_serialization:
|
| 580 |
+
save_file(shard, output_path, metadata={"format": "pt"})
|
| 581 |
+
else:
|
| 582 |
+
torch.save(shard, output_path)
|
| 583 |
+
|
| 584 |
+
# Save index if sharded
|
| 585 |
+
if state_dict_split.is_sharded:
|
| 586 |
+
index = {
|
| 587 |
+
"metadata": state_dict_split.metadata,
|
| 588 |
+
"weight_map": state_dict_split.tensor_to_filename,
|
| 589 |
+
}
|
| 590 |
+
save_index_file = "model.safetensors.index.json" if safe_serialization else "pytorch_model.bin.index.json"
|
| 591 |
+
save_index_file = os.path.join(output_dir, save_index_file)
|
| 592 |
+
with open(save_index_file, "w", encoding="utf-8") as f:
|
| 593 |
+
content = json.dumps(index, indent=2, sort_keys=True) + "\n"
|
| 594 |
+
f.write(content)
|
| 595 |
+
|
| 596 |
+
|
| 597 |
+
def load_state_dict_from_zero_checkpoint(model, checkpoint_dir, tag=None):
|
| 598 |
+
"""
|
| 599 |
+
1. Put the provided model to cpu
|
| 600 |
+
2. Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict``
|
| 601 |
+
3. Load it into the provided model
|
| 602 |
+
|
| 603 |
+
Args:
|
| 604 |
+
- ``model``: the model object to update
|
| 605 |
+
- ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
|
| 606 |
+
- ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
|
| 607 |
+
|
| 608 |
+
Returns:
|
| 609 |
+
- ``model`: modified model
|
| 610 |
+
|
| 611 |
+
Make sure you have plenty of CPU memory available before you call this function. If you don't
|
| 612 |
+
have enough use the ``zero_to_fp32.py`` utility to do the conversion. You will find it
|
| 613 |
+
conveniently placed for you in the checkpoint folder.
|
| 614 |
+
|
| 615 |
+
A typical usage might be ::
|
| 616 |
+
|
| 617 |
+
from deepspeed.utils.zero_to_fp32 import load_state_dict_from_zero_checkpoint
|
| 618 |
+
model = load_state_dict_from_zero_checkpoint(trainer.model, checkpoint_dir)
|
| 619 |
+
# submit to model hub or save the model to share with others
|
| 620 |
+
|
| 621 |
+
Note, that once this was run, the ``model`` will no longer be usable in the deepspeed context
|
| 622 |
+
of the same application. i.e. you will need to re-initialize the deepspeed engine, since
|
| 623 |
+
``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
|
| 624 |
+
|
| 625 |
+
"""
|
| 626 |
+
logger.info(f"Extracting fp32 weights")
|
| 627 |
+
state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag)
|
| 628 |
+
|
| 629 |
+
logger.info(f"Overwriting model with fp32 weights")
|
| 630 |
+
model = model.cpu()
|
| 631 |
+
model.load_state_dict(state_dict, strict=False)
|
| 632 |
+
|
| 633 |
+
return model
|
| 634 |
+
|
| 635 |
+
|
| 636 |
+
if __name__ == "__main__":
|
| 637 |
+
parser = argparse.ArgumentParser()
|
| 638 |
+
parser.add_argument("checkpoint_dir",
|
| 639 |
+
type=str,
|
| 640 |
+
help="path to the desired checkpoint folder, e.g., path/checkpoint-12")
|
| 641 |
+
parser.add_argument("output_dir",
|
| 642 |
+
type=str,
|
| 643 |
+
help="directory to the pytorch fp32 state_dict output files"
|
| 644 |
+
"(e.g. path/checkpoint-12-output/)")
|
| 645 |
+
parser.add_argument(
|
| 646 |
+
"--max_shard_size",
|
| 647 |
+
type=str,
|
| 648 |
+
default="5GB",
|
| 649 |
+
help="The maximum size for a checkpoint before being sharded. Checkpoints shard will then be each of size"
|
| 650 |
+
"lower than this size. If expressed as a string, needs to be digits followed by a unit (like `5MB`"
|
| 651 |
+
"We default it to 5GB in order for models to be able to run easily on free-tier google colab instances"
|
| 652 |
+
"without CPU OOM issues.")
|
| 653 |
+
parser.add_argument(
|
| 654 |
+
"--safe_serialization",
|
| 655 |
+
default=False,
|
| 656 |
+
action='store_true',
|
| 657 |
+
help="Whether to save the model using `safetensors` or the traditional PyTorch way (that uses `pickle`).")
|
| 658 |
+
parser.add_argument("-t",
|
| 659 |
+
"--tag",
|
| 660 |
+
type=str,
|
| 661 |
+
default=None,
|
| 662 |
+
help="checkpoint tag used as a unique identifier for checkpoint. e.g., global_step1")
|
| 663 |
+
parser.add_argument("--exclude_frozen_parameters", action='store_true', help="exclude frozen parameters")
|
| 664 |
+
parser.add_argument("-d", "--debug", action='store_true', help="enable debug")
|
| 665 |
+
args = parser.parse_args()
|
| 666 |
+
|
| 667 |
+
debug = args.debug
|
| 668 |
+
|
| 669 |
+
convert_zero_checkpoint_to_fp32_state_dict(args.checkpoint_dir,
|
| 670 |
+
args.output_dir,
|
| 671 |
+
max_shard_size=args.max_shard_size,
|
| 672 |
+
safe_serialization=args.safe_serialization,
|
| 673 |
+
tag=args.tag,
|
| 674 |
+
exclude_frozen_parameters=args.exclude_frozen_parameters)
|