Upload checkpoint from unmask_tags_math_self_distill_INP_trefleave_last_step_u0.001-1.0_gold1_target1_ce0.0
Browse filesThis view is limited to 50 files because it contains too many changes. See raw diff
- .gitattributes +1 -0
- math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/README.md +202 -0
- math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/adapter_config.json +39 -0
- math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/adapter_model.safetensors +3 -0
- math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-100/README.md +202 -0
- math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-100/adapter_config.json +39 -0
- math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-100/adapter_model.safetensors +3 -0
- math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-100/optimizer.pt +3 -0
- math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-100/rng_state_0.pth +3 -0
- math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-100/rng_state_1.pth +3 -0
- math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-100/scheduler.pt +3 -0
- math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-100/trainer_state.json +293 -0
- math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-100/training_args.bin +3 -0
- math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-1000/README.md +202 -0
- math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-1000/adapter_config.json +39 -0
- math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-1000/adapter_model.safetensors +3 -0
- math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-1000/optimizer.pt +3 -0
- math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-1000/rng_state_0.pth +3 -0
- math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-1000/rng_state_1.pth +3 -0
- math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-1000/scheduler.pt +3 -0
- math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-1000/trainer_state.json +2633 -0
- math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-1000/training_args.bin +3 -0
- math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-1100/README.md +202 -0
- math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-1100/adapter_config.json +39 -0
- math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-1100/adapter_model.safetensors +3 -0
- math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-1100/optimizer.pt +3 -0
- math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-1100/rng_state_0.pth +3 -0
- math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-1100/rng_state_1.pth +3 -0
- math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-1100/scheduler.pt +3 -0
- math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-1100/trainer_state.json +2893 -0
- math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-1100/training_args.bin +3 -0
- math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-1200/README.md +202 -0
- math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-1200/adapter_config.json +39 -0
- math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-1200/adapter_model.safetensors +3 -0
- math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-1200/optimizer.pt +3 -0
- math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-1200/rng_state_0.pth +3 -0
- math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-1200/rng_state_1.pth +3 -0
- math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-1200/scheduler.pt +3 -0
- math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-1200/trainer_state.json +0 -0
- math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-1200/training_args.bin +3 -0
- math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-1300/README.md +202 -0
- math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-1300/adapter_config.json +39 -0
- math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-1300/adapter_model.safetensors +3 -0
- math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-1300/optimizer.pt +3 -0
- math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-1300/rng_state_0.pth +3 -0
- math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-1300/rng_state_1.pth +3 -0
- math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-1300/scheduler.pt +3 -0
- math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-1300/trainer_state.json +0 -0
- math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-1300/training_args.bin +3 -0
- math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-1400/README.md +202 -0
.gitattributes
CHANGED
|
@@ -44,3 +44,4 @@ math/INP-PAR/unmask_tags_gold1_target1_ce0.0/debug_training_examples.jsonl filte
|
|
| 44 |
math/INP-OH/unmask_tags_gold1_target1_ce0.0/debug_training_examples.jsonl filter=lfs diff=lfs merge=lfs -text
|
| 45 |
math/INP/unmask_tags_gold1_target1_ce0.0/debug_training_examples.jsonl filter=lfs diff=lfs merge=lfs -text
|
| 46 |
math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.5/debug_training_examples.jsonl filter=lfs diff=lfs merge=lfs -text
|
|
|
|
|
|
| 44 |
math/INP-OH/unmask_tags_gold1_target1_ce0.0/debug_training_examples.jsonl filter=lfs diff=lfs merge=lfs -text
|
| 45 |
math/INP/unmask_tags_gold1_target1_ce0.0/debug_training_examples.jsonl filter=lfs diff=lfs merge=lfs -text
|
| 46 |
math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.5/debug_training_examples.jsonl filter=lfs diff=lfs merge=lfs -text
|
| 47 |
+
math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/debug_training_examples.jsonl filter=lfs diff=lfs merge=lfs -text
|
math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/README.md
ADDED
|
@@ -0,0 +1,202 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
base_model: GSAI-ML/LLaDA-8B-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 0.15.1
|
math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/adapter_config.json
ADDED
|
@@ -0,0 +1,39 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"alpha_pattern": {},
|
| 3 |
+
"auto_mapping": null,
|
| 4 |
+
"base_model_name_or_path": "GSAI-ML/LLaDA-8B-Instruct",
|
| 5 |
+
"bias": "none",
|
| 6 |
+
"corda_config": null,
|
| 7 |
+
"eva_config": null,
|
| 8 |
+
"exclude_modules": null,
|
| 9 |
+
"fan_in_fan_out": false,
|
| 10 |
+
"inference_mode": true,
|
| 11 |
+
"init_lora_weights": true,
|
| 12 |
+
"layer_replication": null,
|
| 13 |
+
"layers_pattern": null,
|
| 14 |
+
"layers_to_transform": null,
|
| 15 |
+
"loftq_config": {},
|
| 16 |
+
"lora_alpha": 64,
|
| 17 |
+
"lora_bias": false,
|
| 18 |
+
"lora_dropout": 0.05,
|
| 19 |
+
"megatron_config": null,
|
| 20 |
+
"megatron_core": "megatron.core",
|
| 21 |
+
"modules_to_save": null,
|
| 22 |
+
"peft_type": "LORA",
|
| 23 |
+
"r": 128,
|
| 24 |
+
"rank_pattern": {},
|
| 25 |
+
"revision": null,
|
| 26 |
+
"target_modules": [
|
| 27 |
+
"q_proj",
|
| 28 |
+
"k_proj",
|
| 29 |
+
"o_proj",
|
| 30 |
+
"up_proj",
|
| 31 |
+
"gate_proj",
|
| 32 |
+
"v_proj",
|
| 33 |
+
"down_proj"
|
| 34 |
+
],
|
| 35 |
+
"task_type": "CAUSAL_LM",
|
| 36 |
+
"trainable_token_indices": null,
|
| 37 |
+
"use_dora": false,
|
| 38 |
+
"use_rslora": false
|
| 39 |
+
}
|
math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/adapter_model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8626119a692165d0e9fd54745fe58ce922d5ffe613a463952e43cc59b522df4b
|
| 3 |
+
size 2406624648
|
math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-100/README.md
ADDED
|
@@ -0,0 +1,202 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
base_model: GSAI-ML/LLaDA-8B-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 0.15.1
|
math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-100/adapter_config.json
ADDED
|
@@ -0,0 +1,39 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"alpha_pattern": {},
|
| 3 |
+
"auto_mapping": null,
|
| 4 |
+
"base_model_name_or_path": "GSAI-ML/LLaDA-8B-Instruct",
|
| 5 |
+
"bias": "none",
|
| 6 |
+
"corda_config": null,
|
| 7 |
+
"eva_config": null,
|
| 8 |
+
"exclude_modules": null,
|
| 9 |
+
"fan_in_fan_out": false,
|
| 10 |
+
"inference_mode": true,
|
| 11 |
+
"init_lora_weights": true,
|
| 12 |
+
"layer_replication": null,
|
| 13 |
+
"layers_pattern": null,
|
| 14 |
+
"layers_to_transform": null,
|
| 15 |
+
"loftq_config": {},
|
| 16 |
+
"lora_alpha": 64,
|
| 17 |
+
"lora_bias": false,
|
| 18 |
+
"lora_dropout": 0.05,
|
| 19 |
+
"megatron_config": null,
|
| 20 |
+
"megatron_core": "megatron.core",
|
| 21 |
+
"modules_to_save": null,
|
| 22 |
+
"peft_type": "LORA",
|
| 23 |
+
"r": 128,
|
| 24 |
+
"rank_pattern": {},
|
| 25 |
+
"revision": null,
|
| 26 |
+
"target_modules": [
|
| 27 |
+
"q_proj",
|
| 28 |
+
"k_proj",
|
| 29 |
+
"o_proj",
|
| 30 |
+
"up_proj",
|
| 31 |
+
"gate_proj",
|
| 32 |
+
"v_proj",
|
| 33 |
+
"down_proj"
|
| 34 |
+
],
|
| 35 |
+
"task_type": "CAUSAL_LM",
|
| 36 |
+
"trainable_token_indices": null,
|
| 37 |
+
"use_dora": false,
|
| 38 |
+
"use_rslora": false
|
| 39 |
+
}
|
math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-100/adapter_model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:74884d515fc2b2db1e3a891b81ccf2344f78440d1b2a0e8627a0230da6f96a06
|
| 3 |
+
size 2406624648
|
math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-100/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7e5dd9be28fd3a8686f09ebffc66720c469e1039946d6338cfa88153dbd44e4b
|
| 3 |
+
size 671304442
|
math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-100/rng_state_0.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6603180b4f7b4f23eb2fefc470a07c3ec6223e2d309190662f43f30d96be9ce5
|
| 3 |
+
size 14512
|
math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-100/rng_state_1.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:27f1021fb57112918a3a6091b09b0ccd50cb071a2324c12ae9afcc9851ee8bd3
|
| 3 |
+
size 14512
|
math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-100/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3d8cdfceac9f7917b978dca661a3b8e04187faea5d5f6bd7b462d61d8234d57f
|
| 3 |
+
size 1064
|
math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-100/trainer_state.json
ADDED
|
@@ -0,0 +1,293 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.21333333333333335,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 100,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"avg_mask_ratio": 0.4931091487989761,
|
| 13 |
+
"avg_response_length": 225.975,
|
| 14 |
+
"avg_student_mask_ratio": 0.4931091487989761,
|
| 15 |
+
"batch_ainp_frac": 0.0,
|
| 16 |
+
"batch_inp_frac": 1.0,
|
| 17 |
+
"batch_inp_oh_frac": 0.0,
|
| 18 |
+
"batch_inp_par_frac": 0.0,
|
| 19 |
+
"batch_inp_par_par_frac": 0.0,
|
| 20 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 21 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 22 |
+
"batch_rl_frac": 0.0,
|
| 23 |
+
"batch_sft_frac": 0.0,
|
| 24 |
+
"batch_soft_sft_frac": 0.0,
|
| 25 |
+
"batch_tf_frac": 0.0,
|
| 26 |
+
"ce_loss": 0.0,
|
| 27 |
+
"epoch": 0.021333333333333333,
|
| 28 |
+
"grad_norm": 0.1953125,
|
| 29 |
+
"kd_loss": 0.537169519480176,
|
| 30 |
+
"learning_rate": 3e-06,
|
| 31 |
+
"loss": 0.5508,
|
| 32 |
+
"masked_tokens": 111.95,
|
| 33 |
+
"mean_t": 0.5145528071501758,
|
| 34 |
+
"step": 10,
|
| 35 |
+
"student_masked_tokens": 111.95
|
| 36 |
+
},
|
| 37 |
+
{
|
| 38 |
+
"avg_mask_ratio": 0.42058031369233506,
|
| 39 |
+
"avg_response_length": 255.2625,
|
| 40 |
+
"avg_student_mask_ratio": 0.42058031369233506,
|
| 41 |
+
"batch_ainp_frac": 0.0,
|
| 42 |
+
"batch_inp_frac": 1.0,
|
| 43 |
+
"batch_inp_oh_frac": 0.0,
|
| 44 |
+
"batch_inp_par_frac": 0.0,
|
| 45 |
+
"batch_inp_par_par_frac": 0.0,
|
| 46 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 47 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 48 |
+
"batch_rl_frac": 0.0,
|
| 49 |
+
"batch_sft_frac": 0.0,
|
| 50 |
+
"batch_soft_sft_frac": 0.0,
|
| 51 |
+
"batch_tf_frac": 0.0,
|
| 52 |
+
"ce_loss": 0.0,
|
| 53 |
+
"epoch": 0.042666666666666665,
|
| 54 |
+
"grad_norm": 0.23828125,
|
| 55 |
+
"kd_loss": 0.44732311753604304,
|
| 56 |
+
"learning_rate": 3e-06,
|
| 57 |
+
"loss": 0.3866,
|
| 58 |
+
"masked_tokens": 97.35,
|
| 59 |
+
"mean_t": 0.43874448732240123,
|
| 60 |
+
"step": 20,
|
| 61 |
+
"student_masked_tokens": 97.35
|
| 62 |
+
},
|
| 63 |
+
{
|
| 64 |
+
"avg_mask_ratio": 0.4538542575784959,
|
| 65 |
+
"avg_response_length": 211.7625,
|
| 66 |
+
"avg_student_mask_ratio": 0.4538542575784959,
|
| 67 |
+
"batch_ainp_frac": 0.0,
|
| 68 |
+
"batch_inp_frac": 1.0,
|
| 69 |
+
"batch_inp_oh_frac": 0.0,
|
| 70 |
+
"batch_inp_par_frac": 0.0,
|
| 71 |
+
"batch_inp_par_par_frac": 0.0,
|
| 72 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 73 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 74 |
+
"batch_rl_frac": 0.0,
|
| 75 |
+
"batch_sft_frac": 0.0,
|
| 76 |
+
"batch_soft_sft_frac": 0.0,
|
| 77 |
+
"batch_tf_frac": 0.0,
|
| 78 |
+
"ce_loss": 0.0,
|
| 79 |
+
"epoch": 0.064,
|
| 80 |
+
"grad_norm": 0.09130859375,
|
| 81 |
+
"kd_loss": 0.5295996250815808,
|
| 82 |
+
"learning_rate": 3e-06,
|
| 83 |
+
"loss": 0.462,
|
| 84 |
+
"masked_tokens": 110.2,
|
| 85 |
+
"mean_t": 0.4803953981841914,
|
| 86 |
+
"step": 30,
|
| 87 |
+
"student_masked_tokens": 110.2
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"avg_mask_ratio": 0.4207469140383182,
|
| 91 |
+
"avg_response_length": 224.125,
|
| 92 |
+
"avg_student_mask_ratio": 0.4207469140383182,
|
| 93 |
+
"batch_ainp_frac": 0.0,
|
| 94 |
+
"batch_inp_frac": 1.0,
|
| 95 |
+
"batch_inp_oh_frac": 0.0,
|
| 96 |
+
"batch_inp_par_frac": 0.0,
|
| 97 |
+
"batch_inp_par_par_frac": 0.0,
|
| 98 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 99 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 100 |
+
"batch_rl_frac": 0.0,
|
| 101 |
+
"batch_sft_frac": 0.0,
|
| 102 |
+
"batch_soft_sft_frac": 0.0,
|
| 103 |
+
"batch_tf_frac": 0.0,
|
| 104 |
+
"ce_loss": 0.0,
|
| 105 |
+
"epoch": 0.08533333333333333,
|
| 106 |
+
"grad_norm": 0.294921875,
|
| 107 |
+
"kd_loss": 0.3314848062681676,
|
| 108 |
+
"learning_rate": 3e-06,
|
| 109 |
+
"loss": 0.3521,
|
| 110 |
+
"masked_tokens": 98.1625,
|
| 111 |
+
"mean_t": 0.4569831106782658,
|
| 112 |
+
"step": 40,
|
| 113 |
+
"student_masked_tokens": 98.1625
|
| 114 |
+
},
|
| 115 |
+
{
|
| 116 |
+
"avg_mask_ratio": 0.43260439952719026,
|
| 117 |
+
"avg_response_length": 207.125,
|
| 118 |
+
"avg_student_mask_ratio": 0.43260439952719026,
|
| 119 |
+
"batch_ainp_frac": 0.0,
|
| 120 |
+
"batch_inp_frac": 1.0,
|
| 121 |
+
"batch_inp_oh_frac": 0.0,
|
| 122 |
+
"batch_inp_par_frac": 0.0,
|
| 123 |
+
"batch_inp_par_par_frac": 0.0,
|
| 124 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 125 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 126 |
+
"batch_rl_frac": 0.0,
|
| 127 |
+
"batch_sft_frac": 0.0,
|
| 128 |
+
"batch_soft_sft_frac": 0.0,
|
| 129 |
+
"batch_tf_frac": 0.0,
|
| 130 |
+
"ce_loss": 0.0,
|
| 131 |
+
"epoch": 0.10666666666666667,
|
| 132 |
+
"grad_norm": 0.1328125,
|
| 133 |
+
"kd_loss": 0.4259288739462136,
|
| 134 |
+
"learning_rate": 3e-06,
|
| 135 |
+
"loss": 0.4221,
|
| 136 |
+
"masked_tokens": 85.05,
|
| 137 |
+
"mean_t": 0.4612453707959503,
|
| 138 |
+
"step": 50,
|
| 139 |
+
"student_masked_tokens": 85.05
|
| 140 |
+
},
|
| 141 |
+
{
|
| 142 |
+
"avg_mask_ratio": 0.46053453313652426,
|
| 143 |
+
"avg_response_length": 251.0875,
|
| 144 |
+
"avg_student_mask_ratio": 0.46053453313652426,
|
| 145 |
+
"batch_ainp_frac": 0.0,
|
| 146 |
+
"batch_inp_frac": 1.0,
|
| 147 |
+
"batch_inp_oh_frac": 0.0,
|
| 148 |
+
"batch_inp_par_frac": 0.0,
|
| 149 |
+
"batch_inp_par_par_frac": 0.0,
|
| 150 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 151 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 152 |
+
"batch_rl_frac": 0.0,
|
| 153 |
+
"batch_sft_frac": 0.0,
|
| 154 |
+
"batch_soft_sft_frac": 0.0,
|
| 155 |
+
"batch_tf_frac": 0.0,
|
| 156 |
+
"ce_loss": 0.0,
|
| 157 |
+
"epoch": 0.128,
|
| 158 |
+
"grad_norm": 0.126953125,
|
| 159 |
+
"kd_loss": 0.38882399915036386,
|
| 160 |
+
"learning_rate": 3e-06,
|
| 161 |
+
"loss": 0.4142,
|
| 162 |
+
"masked_tokens": 120.9,
|
| 163 |
+
"mean_t": 0.48597636765334756,
|
| 164 |
+
"step": 60,
|
| 165 |
+
"student_masked_tokens": 120.9
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"avg_mask_ratio": 0.5112146578729153,
|
| 169 |
+
"avg_response_length": 202.5875,
|
| 170 |
+
"avg_student_mask_ratio": 0.5112146578729153,
|
| 171 |
+
"batch_ainp_frac": 0.0,
|
| 172 |
+
"batch_inp_frac": 1.0,
|
| 173 |
+
"batch_inp_oh_frac": 0.0,
|
| 174 |
+
"batch_inp_par_frac": 0.0,
|
| 175 |
+
"batch_inp_par_par_frac": 0.0,
|
| 176 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 177 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 178 |
+
"batch_rl_frac": 0.0,
|
| 179 |
+
"batch_sft_frac": 0.0,
|
| 180 |
+
"batch_soft_sft_frac": 0.0,
|
| 181 |
+
"batch_tf_frac": 0.0,
|
| 182 |
+
"ce_loss": 0.0,
|
| 183 |
+
"epoch": 0.14933333333333335,
|
| 184 |
+
"grad_norm": 0.3671875,
|
| 185 |
+
"kd_loss": 0.4423349894936791,
|
| 186 |
+
"learning_rate": 3e-06,
|
| 187 |
+
"loss": 0.4892,
|
| 188 |
+
"masked_tokens": 104.5875,
|
| 189 |
+
"mean_t": 0.5459650319069624,
|
| 190 |
+
"step": 70,
|
| 191 |
+
"student_masked_tokens": 104.5875
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"avg_mask_ratio": 0.37548826879356056,
|
| 195 |
+
"avg_response_length": 225.85,
|
| 196 |
+
"avg_student_mask_ratio": 0.37548826879356056,
|
| 197 |
+
"batch_ainp_frac": 0.0,
|
| 198 |
+
"batch_inp_frac": 1.0,
|
| 199 |
+
"batch_inp_oh_frac": 0.0,
|
| 200 |
+
"batch_inp_par_frac": 0.0,
|
| 201 |
+
"batch_inp_par_par_frac": 0.0,
|
| 202 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 203 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 204 |
+
"batch_rl_frac": 0.0,
|
| 205 |
+
"batch_sft_frac": 0.0,
|
| 206 |
+
"batch_soft_sft_frac": 0.0,
|
| 207 |
+
"batch_tf_frac": 0.0,
|
| 208 |
+
"ce_loss": 0.0,
|
| 209 |
+
"epoch": 0.17066666666666666,
|
| 210 |
+
"grad_norm": 0.083984375,
|
| 211 |
+
"kd_loss": 0.3107750686068783,
|
| 212 |
+
"learning_rate": 3e-06,
|
| 213 |
+
"loss": 0.3183,
|
| 214 |
+
"masked_tokens": 85.0625,
|
| 215 |
+
"mean_t": 0.40758824030635876,
|
| 216 |
+
"step": 80,
|
| 217 |
+
"student_masked_tokens": 85.0625
|
| 218 |
+
},
|
| 219 |
+
{
|
| 220 |
+
"avg_mask_ratio": 0.5001560213277116,
|
| 221 |
+
"avg_response_length": 229.75,
|
| 222 |
+
"avg_student_mask_ratio": 0.5001560213277116,
|
| 223 |
+
"batch_ainp_frac": 0.0,
|
| 224 |
+
"batch_inp_frac": 1.0,
|
| 225 |
+
"batch_inp_oh_frac": 0.0,
|
| 226 |
+
"batch_inp_par_frac": 0.0,
|
| 227 |
+
"batch_inp_par_par_frac": 0.0,
|
| 228 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 229 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 230 |
+
"batch_rl_frac": 0.0,
|
| 231 |
+
"batch_sft_frac": 0.0,
|
| 232 |
+
"batch_soft_sft_frac": 0.0,
|
| 233 |
+
"batch_tf_frac": 0.0,
|
| 234 |
+
"ce_loss": 0.0,
|
| 235 |
+
"epoch": 0.192,
|
| 236 |
+
"grad_norm": 0.57421875,
|
| 237 |
+
"kd_loss": 0.600051444793823,
|
| 238 |
+
"learning_rate": 3e-06,
|
| 239 |
+
"loss": 0.578,
|
| 240 |
+
"masked_tokens": 107.6625,
|
| 241 |
+
"mean_t": 0.5297661645396147,
|
| 242 |
+
"step": 90,
|
| 243 |
+
"student_masked_tokens": 107.6625
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"avg_mask_ratio": 0.4576045103633078,
|
| 247 |
+
"avg_response_length": 208.0,
|
| 248 |
+
"avg_student_mask_ratio": 0.4576045103633078,
|
| 249 |
+
"batch_ainp_frac": 0.0,
|
| 250 |
+
"batch_inp_frac": 1.0,
|
| 251 |
+
"batch_inp_oh_frac": 0.0,
|
| 252 |
+
"batch_inp_par_frac": 0.0,
|
| 253 |
+
"batch_inp_par_par_frac": 0.0,
|
| 254 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 255 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 256 |
+
"batch_rl_frac": 0.0,
|
| 257 |
+
"batch_sft_frac": 0.0,
|
| 258 |
+
"batch_soft_sft_frac": 0.0,
|
| 259 |
+
"batch_tf_frac": 0.0,
|
| 260 |
+
"ce_loss": 0.0,
|
| 261 |
+
"epoch": 0.21333333333333335,
|
| 262 |
+
"grad_norm": 0.1767578125,
|
| 263 |
+
"kd_loss": 0.38177184401774583,
|
| 264 |
+
"learning_rate": 3e-06,
|
| 265 |
+
"loss": 0.4107,
|
| 266 |
+
"masked_tokens": 104.1625,
|
| 267 |
+
"mean_t": 0.4886587227345444,
|
| 268 |
+
"step": 100,
|
| 269 |
+
"student_masked_tokens": 104.1625
|
| 270 |
+
}
|
| 271 |
+
],
|
| 272 |
+
"logging_steps": 10,
|
| 273 |
+
"max_steps": 1404,
|
| 274 |
+
"num_input_tokens_seen": 0,
|
| 275 |
+
"num_train_epochs": 3,
|
| 276 |
+
"save_steps": 100,
|
| 277 |
+
"stateful_callbacks": {
|
| 278 |
+
"TrainerControl": {
|
| 279 |
+
"args": {
|
| 280 |
+
"should_epoch_stop": false,
|
| 281 |
+
"should_evaluate": false,
|
| 282 |
+
"should_log": false,
|
| 283 |
+
"should_save": true,
|
| 284 |
+
"should_training_stop": false
|
| 285 |
+
},
|
| 286 |
+
"attributes": {}
|
| 287 |
+
}
|
| 288 |
+
},
|
| 289 |
+
"total_flos": 0.0,
|
| 290 |
+
"train_batch_size": 1,
|
| 291 |
+
"trial_name": null,
|
| 292 |
+
"trial_params": null
|
| 293 |
+
}
|
math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-100/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:99e6e32f3c28324ecee9c50eb13825addacaf6ce31b5b6707966fc1ac725dea3
|
| 3 |
+
size 8312
|
math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-1000/README.md
ADDED
|
@@ -0,0 +1,202 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
base_model: GSAI-ML/LLaDA-8B-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 0.15.1
|
math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-1000/adapter_config.json
ADDED
|
@@ -0,0 +1,39 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"alpha_pattern": {},
|
| 3 |
+
"auto_mapping": null,
|
| 4 |
+
"base_model_name_or_path": "GSAI-ML/LLaDA-8B-Instruct",
|
| 5 |
+
"bias": "none",
|
| 6 |
+
"corda_config": null,
|
| 7 |
+
"eva_config": null,
|
| 8 |
+
"exclude_modules": null,
|
| 9 |
+
"fan_in_fan_out": false,
|
| 10 |
+
"inference_mode": true,
|
| 11 |
+
"init_lora_weights": true,
|
| 12 |
+
"layer_replication": null,
|
| 13 |
+
"layers_pattern": null,
|
| 14 |
+
"layers_to_transform": null,
|
| 15 |
+
"loftq_config": {},
|
| 16 |
+
"lora_alpha": 64,
|
| 17 |
+
"lora_bias": false,
|
| 18 |
+
"lora_dropout": 0.05,
|
| 19 |
+
"megatron_config": null,
|
| 20 |
+
"megatron_core": "megatron.core",
|
| 21 |
+
"modules_to_save": null,
|
| 22 |
+
"peft_type": "LORA",
|
| 23 |
+
"r": 128,
|
| 24 |
+
"rank_pattern": {},
|
| 25 |
+
"revision": null,
|
| 26 |
+
"target_modules": [
|
| 27 |
+
"q_proj",
|
| 28 |
+
"k_proj",
|
| 29 |
+
"o_proj",
|
| 30 |
+
"up_proj",
|
| 31 |
+
"gate_proj",
|
| 32 |
+
"v_proj",
|
| 33 |
+
"down_proj"
|
| 34 |
+
],
|
| 35 |
+
"task_type": "CAUSAL_LM",
|
| 36 |
+
"trainable_token_indices": null,
|
| 37 |
+
"use_dora": false,
|
| 38 |
+
"use_rslora": false
|
| 39 |
+
}
|
math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-1000/adapter_model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:bd2b63dcdc8d6b6cf9817f8a4bbc739d122c42992edfc8102506690521ada5b7
|
| 3 |
+
size 2406624648
|
math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-1000/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:77e7e0eb60c1e815f368c79c8fb0cc87e6a387cfba46e28508d9feb975f351cf
|
| 3 |
+
size 671304442
|
math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-1000/rng_state_0.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6cbbccc533dc6035b9eb3e81ab0c37a3544ee2638528b1cb900a84d35f5b76b2
|
| 3 |
+
size 14512
|
math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-1000/rng_state_1.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:69989ab343a15b03b088a3c9e2bb39b88ed718b675223b5e1d55890e63636453
|
| 3 |
+
size 14512
|
math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-1000/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:29d9aa99505fc60c0db1b9cdacaa08b06e8a85c8aaaab4e389667a719fafb9bf
|
| 3 |
+
size 1064
|
math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-1000/trainer_state.json
ADDED
|
@@ -0,0 +1,2633 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 2.1365333333333334,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 1000,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"avg_mask_ratio": 0.4931091487989761,
|
| 13 |
+
"avg_response_length": 225.975,
|
| 14 |
+
"avg_student_mask_ratio": 0.4931091487989761,
|
| 15 |
+
"batch_ainp_frac": 0.0,
|
| 16 |
+
"batch_inp_frac": 1.0,
|
| 17 |
+
"batch_inp_oh_frac": 0.0,
|
| 18 |
+
"batch_inp_par_frac": 0.0,
|
| 19 |
+
"batch_inp_par_par_frac": 0.0,
|
| 20 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 21 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 22 |
+
"batch_rl_frac": 0.0,
|
| 23 |
+
"batch_sft_frac": 0.0,
|
| 24 |
+
"batch_soft_sft_frac": 0.0,
|
| 25 |
+
"batch_tf_frac": 0.0,
|
| 26 |
+
"ce_loss": 0.0,
|
| 27 |
+
"epoch": 0.021333333333333333,
|
| 28 |
+
"grad_norm": 0.1953125,
|
| 29 |
+
"kd_loss": 0.537169519480176,
|
| 30 |
+
"learning_rate": 3e-06,
|
| 31 |
+
"loss": 0.5508,
|
| 32 |
+
"masked_tokens": 111.95,
|
| 33 |
+
"mean_t": 0.5145528071501758,
|
| 34 |
+
"step": 10,
|
| 35 |
+
"student_masked_tokens": 111.95
|
| 36 |
+
},
|
| 37 |
+
{
|
| 38 |
+
"avg_mask_ratio": 0.42058031369233506,
|
| 39 |
+
"avg_response_length": 255.2625,
|
| 40 |
+
"avg_student_mask_ratio": 0.42058031369233506,
|
| 41 |
+
"batch_ainp_frac": 0.0,
|
| 42 |
+
"batch_inp_frac": 1.0,
|
| 43 |
+
"batch_inp_oh_frac": 0.0,
|
| 44 |
+
"batch_inp_par_frac": 0.0,
|
| 45 |
+
"batch_inp_par_par_frac": 0.0,
|
| 46 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 47 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 48 |
+
"batch_rl_frac": 0.0,
|
| 49 |
+
"batch_sft_frac": 0.0,
|
| 50 |
+
"batch_soft_sft_frac": 0.0,
|
| 51 |
+
"batch_tf_frac": 0.0,
|
| 52 |
+
"ce_loss": 0.0,
|
| 53 |
+
"epoch": 0.042666666666666665,
|
| 54 |
+
"grad_norm": 0.23828125,
|
| 55 |
+
"kd_loss": 0.44732311753604304,
|
| 56 |
+
"learning_rate": 3e-06,
|
| 57 |
+
"loss": 0.3866,
|
| 58 |
+
"masked_tokens": 97.35,
|
| 59 |
+
"mean_t": 0.43874448732240123,
|
| 60 |
+
"step": 20,
|
| 61 |
+
"student_masked_tokens": 97.35
|
| 62 |
+
},
|
| 63 |
+
{
|
| 64 |
+
"avg_mask_ratio": 0.4538542575784959,
|
| 65 |
+
"avg_response_length": 211.7625,
|
| 66 |
+
"avg_student_mask_ratio": 0.4538542575784959,
|
| 67 |
+
"batch_ainp_frac": 0.0,
|
| 68 |
+
"batch_inp_frac": 1.0,
|
| 69 |
+
"batch_inp_oh_frac": 0.0,
|
| 70 |
+
"batch_inp_par_frac": 0.0,
|
| 71 |
+
"batch_inp_par_par_frac": 0.0,
|
| 72 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 73 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 74 |
+
"batch_rl_frac": 0.0,
|
| 75 |
+
"batch_sft_frac": 0.0,
|
| 76 |
+
"batch_soft_sft_frac": 0.0,
|
| 77 |
+
"batch_tf_frac": 0.0,
|
| 78 |
+
"ce_loss": 0.0,
|
| 79 |
+
"epoch": 0.064,
|
| 80 |
+
"grad_norm": 0.09130859375,
|
| 81 |
+
"kd_loss": 0.5295996250815808,
|
| 82 |
+
"learning_rate": 3e-06,
|
| 83 |
+
"loss": 0.462,
|
| 84 |
+
"masked_tokens": 110.2,
|
| 85 |
+
"mean_t": 0.4803953981841914,
|
| 86 |
+
"step": 30,
|
| 87 |
+
"student_masked_tokens": 110.2
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"avg_mask_ratio": 0.4207469140383182,
|
| 91 |
+
"avg_response_length": 224.125,
|
| 92 |
+
"avg_student_mask_ratio": 0.4207469140383182,
|
| 93 |
+
"batch_ainp_frac": 0.0,
|
| 94 |
+
"batch_inp_frac": 1.0,
|
| 95 |
+
"batch_inp_oh_frac": 0.0,
|
| 96 |
+
"batch_inp_par_frac": 0.0,
|
| 97 |
+
"batch_inp_par_par_frac": 0.0,
|
| 98 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 99 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 100 |
+
"batch_rl_frac": 0.0,
|
| 101 |
+
"batch_sft_frac": 0.0,
|
| 102 |
+
"batch_soft_sft_frac": 0.0,
|
| 103 |
+
"batch_tf_frac": 0.0,
|
| 104 |
+
"ce_loss": 0.0,
|
| 105 |
+
"epoch": 0.08533333333333333,
|
| 106 |
+
"grad_norm": 0.294921875,
|
| 107 |
+
"kd_loss": 0.3314848062681676,
|
| 108 |
+
"learning_rate": 3e-06,
|
| 109 |
+
"loss": 0.3521,
|
| 110 |
+
"masked_tokens": 98.1625,
|
| 111 |
+
"mean_t": 0.4569831106782658,
|
| 112 |
+
"step": 40,
|
| 113 |
+
"student_masked_tokens": 98.1625
|
| 114 |
+
},
|
| 115 |
+
{
|
| 116 |
+
"avg_mask_ratio": 0.43260439952719026,
|
| 117 |
+
"avg_response_length": 207.125,
|
| 118 |
+
"avg_student_mask_ratio": 0.43260439952719026,
|
| 119 |
+
"batch_ainp_frac": 0.0,
|
| 120 |
+
"batch_inp_frac": 1.0,
|
| 121 |
+
"batch_inp_oh_frac": 0.0,
|
| 122 |
+
"batch_inp_par_frac": 0.0,
|
| 123 |
+
"batch_inp_par_par_frac": 0.0,
|
| 124 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 125 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 126 |
+
"batch_rl_frac": 0.0,
|
| 127 |
+
"batch_sft_frac": 0.0,
|
| 128 |
+
"batch_soft_sft_frac": 0.0,
|
| 129 |
+
"batch_tf_frac": 0.0,
|
| 130 |
+
"ce_loss": 0.0,
|
| 131 |
+
"epoch": 0.10666666666666667,
|
| 132 |
+
"grad_norm": 0.1328125,
|
| 133 |
+
"kd_loss": 0.4259288739462136,
|
| 134 |
+
"learning_rate": 3e-06,
|
| 135 |
+
"loss": 0.4221,
|
| 136 |
+
"masked_tokens": 85.05,
|
| 137 |
+
"mean_t": 0.4612453707959503,
|
| 138 |
+
"step": 50,
|
| 139 |
+
"student_masked_tokens": 85.05
|
| 140 |
+
},
|
| 141 |
+
{
|
| 142 |
+
"avg_mask_ratio": 0.46053453313652426,
|
| 143 |
+
"avg_response_length": 251.0875,
|
| 144 |
+
"avg_student_mask_ratio": 0.46053453313652426,
|
| 145 |
+
"batch_ainp_frac": 0.0,
|
| 146 |
+
"batch_inp_frac": 1.0,
|
| 147 |
+
"batch_inp_oh_frac": 0.0,
|
| 148 |
+
"batch_inp_par_frac": 0.0,
|
| 149 |
+
"batch_inp_par_par_frac": 0.0,
|
| 150 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 151 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 152 |
+
"batch_rl_frac": 0.0,
|
| 153 |
+
"batch_sft_frac": 0.0,
|
| 154 |
+
"batch_soft_sft_frac": 0.0,
|
| 155 |
+
"batch_tf_frac": 0.0,
|
| 156 |
+
"ce_loss": 0.0,
|
| 157 |
+
"epoch": 0.128,
|
| 158 |
+
"grad_norm": 0.126953125,
|
| 159 |
+
"kd_loss": 0.38882399915036386,
|
| 160 |
+
"learning_rate": 3e-06,
|
| 161 |
+
"loss": 0.4142,
|
| 162 |
+
"masked_tokens": 120.9,
|
| 163 |
+
"mean_t": 0.48597636765334756,
|
| 164 |
+
"step": 60,
|
| 165 |
+
"student_masked_tokens": 120.9
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"avg_mask_ratio": 0.5112146578729153,
|
| 169 |
+
"avg_response_length": 202.5875,
|
| 170 |
+
"avg_student_mask_ratio": 0.5112146578729153,
|
| 171 |
+
"batch_ainp_frac": 0.0,
|
| 172 |
+
"batch_inp_frac": 1.0,
|
| 173 |
+
"batch_inp_oh_frac": 0.0,
|
| 174 |
+
"batch_inp_par_frac": 0.0,
|
| 175 |
+
"batch_inp_par_par_frac": 0.0,
|
| 176 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 177 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 178 |
+
"batch_rl_frac": 0.0,
|
| 179 |
+
"batch_sft_frac": 0.0,
|
| 180 |
+
"batch_soft_sft_frac": 0.0,
|
| 181 |
+
"batch_tf_frac": 0.0,
|
| 182 |
+
"ce_loss": 0.0,
|
| 183 |
+
"epoch": 0.14933333333333335,
|
| 184 |
+
"grad_norm": 0.3671875,
|
| 185 |
+
"kd_loss": 0.4423349894936791,
|
| 186 |
+
"learning_rate": 3e-06,
|
| 187 |
+
"loss": 0.4892,
|
| 188 |
+
"masked_tokens": 104.5875,
|
| 189 |
+
"mean_t": 0.5459650319069624,
|
| 190 |
+
"step": 70,
|
| 191 |
+
"student_masked_tokens": 104.5875
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"avg_mask_ratio": 0.37548826879356056,
|
| 195 |
+
"avg_response_length": 225.85,
|
| 196 |
+
"avg_student_mask_ratio": 0.37548826879356056,
|
| 197 |
+
"batch_ainp_frac": 0.0,
|
| 198 |
+
"batch_inp_frac": 1.0,
|
| 199 |
+
"batch_inp_oh_frac": 0.0,
|
| 200 |
+
"batch_inp_par_frac": 0.0,
|
| 201 |
+
"batch_inp_par_par_frac": 0.0,
|
| 202 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 203 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 204 |
+
"batch_rl_frac": 0.0,
|
| 205 |
+
"batch_sft_frac": 0.0,
|
| 206 |
+
"batch_soft_sft_frac": 0.0,
|
| 207 |
+
"batch_tf_frac": 0.0,
|
| 208 |
+
"ce_loss": 0.0,
|
| 209 |
+
"epoch": 0.17066666666666666,
|
| 210 |
+
"grad_norm": 0.083984375,
|
| 211 |
+
"kd_loss": 0.3107750686068783,
|
| 212 |
+
"learning_rate": 3e-06,
|
| 213 |
+
"loss": 0.3183,
|
| 214 |
+
"masked_tokens": 85.0625,
|
| 215 |
+
"mean_t": 0.40758824030635876,
|
| 216 |
+
"step": 80,
|
| 217 |
+
"student_masked_tokens": 85.0625
|
| 218 |
+
},
|
| 219 |
+
{
|
| 220 |
+
"avg_mask_ratio": 0.5001560213277116,
|
| 221 |
+
"avg_response_length": 229.75,
|
| 222 |
+
"avg_student_mask_ratio": 0.5001560213277116,
|
| 223 |
+
"batch_ainp_frac": 0.0,
|
| 224 |
+
"batch_inp_frac": 1.0,
|
| 225 |
+
"batch_inp_oh_frac": 0.0,
|
| 226 |
+
"batch_inp_par_frac": 0.0,
|
| 227 |
+
"batch_inp_par_par_frac": 0.0,
|
| 228 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 229 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 230 |
+
"batch_rl_frac": 0.0,
|
| 231 |
+
"batch_sft_frac": 0.0,
|
| 232 |
+
"batch_soft_sft_frac": 0.0,
|
| 233 |
+
"batch_tf_frac": 0.0,
|
| 234 |
+
"ce_loss": 0.0,
|
| 235 |
+
"epoch": 0.192,
|
| 236 |
+
"grad_norm": 0.57421875,
|
| 237 |
+
"kd_loss": 0.600051444793823,
|
| 238 |
+
"learning_rate": 3e-06,
|
| 239 |
+
"loss": 0.578,
|
| 240 |
+
"masked_tokens": 107.6625,
|
| 241 |
+
"mean_t": 0.5297661645396147,
|
| 242 |
+
"step": 90,
|
| 243 |
+
"student_masked_tokens": 107.6625
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"avg_mask_ratio": 0.4576045103633078,
|
| 247 |
+
"avg_response_length": 208.0,
|
| 248 |
+
"avg_student_mask_ratio": 0.4576045103633078,
|
| 249 |
+
"batch_ainp_frac": 0.0,
|
| 250 |
+
"batch_inp_frac": 1.0,
|
| 251 |
+
"batch_inp_oh_frac": 0.0,
|
| 252 |
+
"batch_inp_par_frac": 0.0,
|
| 253 |
+
"batch_inp_par_par_frac": 0.0,
|
| 254 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 255 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 256 |
+
"batch_rl_frac": 0.0,
|
| 257 |
+
"batch_sft_frac": 0.0,
|
| 258 |
+
"batch_soft_sft_frac": 0.0,
|
| 259 |
+
"batch_tf_frac": 0.0,
|
| 260 |
+
"ce_loss": 0.0,
|
| 261 |
+
"epoch": 0.21333333333333335,
|
| 262 |
+
"grad_norm": 0.1767578125,
|
| 263 |
+
"kd_loss": 0.38177184401774583,
|
| 264 |
+
"learning_rate": 3e-06,
|
| 265 |
+
"loss": 0.4107,
|
| 266 |
+
"masked_tokens": 104.1625,
|
| 267 |
+
"mean_t": 0.4886587227345444,
|
| 268 |
+
"step": 100,
|
| 269 |
+
"student_masked_tokens": 104.1625
|
| 270 |
+
},
|
| 271 |
+
{
|
| 272 |
+
"avg_mask_ratio": 0.4877026333590038,
|
| 273 |
+
"avg_response_length": 213.0875,
|
| 274 |
+
"avg_student_mask_ratio": 0.4877026333590038,
|
| 275 |
+
"batch_ainp_frac": 0.0,
|
| 276 |
+
"batch_inp_frac": 1.0,
|
| 277 |
+
"batch_inp_oh_frac": 0.0,
|
| 278 |
+
"batch_inp_par_frac": 0.0,
|
| 279 |
+
"batch_inp_par_par_frac": 0.0,
|
| 280 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 281 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 282 |
+
"batch_rl_frac": 0.0,
|
| 283 |
+
"batch_sft_frac": 0.0,
|
| 284 |
+
"batch_soft_sft_frac": 0.0,
|
| 285 |
+
"batch_tf_frac": 0.0,
|
| 286 |
+
"ce_loss": 0.0,
|
| 287 |
+
"epoch": 0.23466666666666666,
|
| 288 |
+
"grad_norm": 0.333984375,
|
| 289 |
+
"kd_loss": 0.50661907401431,
|
| 290 |
+
"learning_rate": 3e-06,
|
| 291 |
+
"loss": 0.5293,
|
| 292 |
+
"masked_tokens": 102.075,
|
| 293 |
+
"mean_t": 0.5246987929102034,
|
| 294 |
+
"step": 110,
|
| 295 |
+
"student_masked_tokens": 102.075
|
| 296 |
+
},
|
| 297 |
+
{
|
| 298 |
+
"avg_mask_ratio": 0.45146879020612685,
|
| 299 |
+
"avg_response_length": 224.1875,
|
| 300 |
+
"avg_student_mask_ratio": 0.45146879020612685,
|
| 301 |
+
"batch_ainp_frac": 0.0,
|
| 302 |
+
"batch_inp_frac": 1.0,
|
| 303 |
+
"batch_inp_oh_frac": 0.0,
|
| 304 |
+
"batch_inp_par_frac": 0.0,
|
| 305 |
+
"batch_inp_par_par_frac": 0.0,
|
| 306 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 307 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 308 |
+
"batch_rl_frac": 0.0,
|
| 309 |
+
"batch_sft_frac": 0.0,
|
| 310 |
+
"batch_soft_sft_frac": 0.0,
|
| 311 |
+
"batch_tf_frac": 0.0,
|
| 312 |
+
"ce_loss": 0.0,
|
| 313 |
+
"epoch": 0.256,
|
| 314 |
+
"grad_norm": 0.236328125,
|
| 315 |
+
"kd_loss": 0.4163589179806877,
|
| 316 |
+
"learning_rate": 3e-06,
|
| 317 |
+
"loss": 0.435,
|
| 318 |
+
"masked_tokens": 100.525,
|
| 319 |
+
"mean_t": 0.4805434140143916,
|
| 320 |
+
"step": 120,
|
| 321 |
+
"student_masked_tokens": 100.525
|
| 322 |
+
},
|
| 323 |
+
{
|
| 324 |
+
"avg_mask_ratio": 0.4356566035945434,
|
| 325 |
+
"avg_response_length": 202.7,
|
| 326 |
+
"avg_student_mask_ratio": 0.4356566035945434,
|
| 327 |
+
"batch_ainp_frac": 0.0,
|
| 328 |
+
"batch_inp_frac": 1.0,
|
| 329 |
+
"batch_inp_oh_frac": 0.0,
|
| 330 |
+
"batch_inp_par_frac": 0.0,
|
| 331 |
+
"batch_inp_par_par_frac": 0.0,
|
| 332 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 333 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 334 |
+
"batch_rl_frac": 0.0,
|
| 335 |
+
"batch_sft_frac": 0.0,
|
| 336 |
+
"batch_soft_sft_frac": 0.0,
|
| 337 |
+
"batch_tf_frac": 0.0,
|
| 338 |
+
"ce_loss": 0.0,
|
| 339 |
+
"epoch": 0.2773333333333333,
|
| 340 |
+
"grad_norm": 0.2314453125,
|
| 341 |
+
"kd_loss": 0.2308679626777348,
|
| 342 |
+
"learning_rate": 3e-06,
|
| 343 |
+
"loss": 0.2829,
|
| 344 |
+
"masked_tokens": 89.5625,
|
| 345 |
+
"mean_t": 0.47522516988683494,
|
| 346 |
+
"step": 130,
|
| 347 |
+
"student_masked_tokens": 89.5625
|
| 348 |
+
},
|
| 349 |
+
{
|
| 350 |
+
"avg_mask_ratio": 0.49419954856857656,
|
| 351 |
+
"avg_response_length": 255.625,
|
| 352 |
+
"avg_student_mask_ratio": 0.49419954856857656,
|
| 353 |
+
"batch_ainp_frac": 0.0,
|
| 354 |
+
"batch_inp_frac": 1.0,
|
| 355 |
+
"batch_inp_oh_frac": 0.0,
|
| 356 |
+
"batch_inp_par_frac": 0.0,
|
| 357 |
+
"batch_inp_par_par_frac": 0.0,
|
| 358 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 359 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 360 |
+
"batch_rl_frac": 0.0,
|
| 361 |
+
"batch_sft_frac": 0.0,
|
| 362 |
+
"batch_soft_sft_frac": 0.0,
|
| 363 |
+
"batch_tf_frac": 0.0,
|
| 364 |
+
"ce_loss": 0.0,
|
| 365 |
+
"epoch": 0.2986666666666667,
|
| 366 |
+
"grad_norm": 0.353515625,
|
| 367 |
+
"kd_loss": 0.5027470885725507,
|
| 368 |
+
"learning_rate": 3e-06,
|
| 369 |
+
"loss": 0.5364,
|
| 370 |
+
"masked_tokens": 136.575,
|
| 371 |
+
"mean_t": 0.5204090005659964,
|
| 372 |
+
"step": 140,
|
| 373 |
+
"student_masked_tokens": 136.575
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"avg_mask_ratio": 0.4736677930341102,
|
| 377 |
+
"avg_response_length": 255.375,
|
| 378 |
+
"avg_student_mask_ratio": 0.4736677930341102,
|
| 379 |
+
"batch_ainp_frac": 0.0,
|
| 380 |
+
"batch_inp_frac": 1.0,
|
| 381 |
+
"batch_inp_oh_frac": 0.0,
|
| 382 |
+
"batch_inp_par_frac": 0.0,
|
| 383 |
+
"batch_inp_par_par_frac": 0.0,
|
| 384 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 385 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 386 |
+
"batch_rl_frac": 0.0,
|
| 387 |
+
"batch_sft_frac": 0.0,
|
| 388 |
+
"batch_soft_sft_frac": 0.0,
|
| 389 |
+
"batch_tf_frac": 0.0,
|
| 390 |
+
"ce_loss": 0.0,
|
| 391 |
+
"epoch": 0.32,
|
| 392 |
+
"grad_norm": 0.07666015625,
|
| 393 |
+
"kd_loss": 0.4891943664209066,
|
| 394 |
+
"learning_rate": 3e-06,
|
| 395 |
+
"loss": 0.5326,
|
| 396 |
+
"masked_tokens": 126.575,
|
| 397 |
+
"mean_t": 0.5012552456930279,
|
| 398 |
+
"step": 150,
|
| 399 |
+
"student_masked_tokens": 126.575
|
| 400 |
+
},
|
| 401 |
+
{
|
| 402 |
+
"avg_mask_ratio": 0.5032523009285796,
|
| 403 |
+
"avg_response_length": 209.325,
|
| 404 |
+
"avg_student_mask_ratio": 0.5032523009285796,
|
| 405 |
+
"batch_ainp_frac": 0.0,
|
| 406 |
+
"batch_inp_frac": 1.0,
|
| 407 |
+
"batch_inp_oh_frac": 0.0,
|
| 408 |
+
"batch_inp_par_frac": 0.0,
|
| 409 |
+
"batch_inp_par_par_frac": 0.0,
|
| 410 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 411 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 412 |
+
"batch_rl_frac": 0.0,
|
| 413 |
+
"batch_sft_frac": 0.0,
|
| 414 |
+
"batch_soft_sft_frac": 0.0,
|
| 415 |
+
"batch_tf_frac": 0.0,
|
| 416 |
+
"ce_loss": 0.0,
|
| 417 |
+
"epoch": 0.3413333333333333,
|
| 418 |
+
"grad_norm": 0.353515625,
|
| 419 |
+
"kd_loss": 0.4466663982801265,
|
| 420 |
+
"learning_rate": 3e-06,
|
| 421 |
+
"loss": 0.487,
|
| 422 |
+
"masked_tokens": 99.425,
|
| 423 |
+
"mean_t": 0.5408745193795766,
|
| 424 |
+
"step": 160,
|
| 425 |
+
"student_masked_tokens": 99.425
|
| 426 |
+
},
|
| 427 |
+
{
|
| 428 |
+
"avg_mask_ratio": 0.4806730231270194,
|
| 429 |
+
"avg_response_length": 190.325,
|
| 430 |
+
"avg_student_mask_ratio": 0.4806730231270194,
|
| 431 |
+
"batch_ainp_frac": 0.0,
|
| 432 |
+
"batch_inp_frac": 1.0,
|
| 433 |
+
"batch_inp_oh_frac": 0.0,
|
| 434 |
+
"batch_inp_par_frac": 0.0,
|
| 435 |
+
"batch_inp_par_par_frac": 0.0,
|
| 436 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 437 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 438 |
+
"batch_rl_frac": 0.0,
|
| 439 |
+
"batch_sft_frac": 0.0,
|
| 440 |
+
"batch_soft_sft_frac": 0.0,
|
| 441 |
+
"batch_tf_frac": 0.0,
|
| 442 |
+
"ce_loss": 0.0,
|
| 443 |
+
"epoch": 0.3626666666666667,
|
| 444 |
+
"grad_norm": 0.1650390625,
|
| 445 |
+
"kd_loss": 0.45357214291389025,
|
| 446 |
+
"learning_rate": 3e-06,
|
| 447 |
+
"loss": 0.4277,
|
| 448 |
+
"masked_tokens": 89.825,
|
| 449 |
+
"mean_t": 0.5134547733236104,
|
| 450 |
+
"step": 170,
|
| 451 |
+
"student_masked_tokens": 89.825
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"avg_mask_ratio": 0.45829249716189224,
|
| 455 |
+
"avg_response_length": 244.0,
|
| 456 |
+
"avg_student_mask_ratio": 0.45829249716189224,
|
| 457 |
+
"batch_ainp_frac": 0.0,
|
| 458 |
+
"batch_inp_frac": 1.0,
|
| 459 |
+
"batch_inp_oh_frac": 0.0,
|
| 460 |
+
"batch_inp_par_frac": 0.0,
|
| 461 |
+
"batch_inp_par_par_frac": 0.0,
|
| 462 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 463 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 464 |
+
"batch_rl_frac": 0.0,
|
| 465 |
+
"batch_sft_frac": 0.0,
|
| 466 |
+
"batch_soft_sft_frac": 0.0,
|
| 467 |
+
"batch_tf_frac": 0.0,
|
| 468 |
+
"ce_loss": 0.0,
|
| 469 |
+
"epoch": 0.384,
|
| 470 |
+
"grad_norm": 0.205078125,
|
| 471 |
+
"kd_loss": 0.35104387890574457,
|
| 472 |
+
"learning_rate": 3e-06,
|
| 473 |
+
"loss": 0.3341,
|
| 474 |
+
"masked_tokens": 110.075,
|
| 475 |
+
"mean_t": 0.48226988823735156,
|
| 476 |
+
"step": 180,
|
| 477 |
+
"student_masked_tokens": 110.075
|
| 478 |
+
},
|
| 479 |
+
{
|
| 480 |
+
"avg_mask_ratio": 0.4842760307248682,
|
| 481 |
+
"avg_response_length": 233.675,
|
| 482 |
+
"avg_student_mask_ratio": 0.4842760307248682,
|
| 483 |
+
"batch_ainp_frac": 0.0,
|
| 484 |
+
"batch_inp_frac": 1.0,
|
| 485 |
+
"batch_inp_oh_frac": 0.0,
|
| 486 |
+
"batch_inp_par_frac": 0.0,
|
| 487 |
+
"batch_inp_par_par_frac": 0.0,
|
| 488 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 489 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 490 |
+
"batch_rl_frac": 0.0,
|
| 491 |
+
"batch_sft_frac": 0.0,
|
| 492 |
+
"batch_soft_sft_frac": 0.0,
|
| 493 |
+
"batch_tf_frac": 0.0,
|
| 494 |
+
"ce_loss": 0.0,
|
| 495 |
+
"epoch": 0.4053333333333333,
|
| 496 |
+
"grad_norm": 0.84375,
|
| 497 |
+
"kd_loss": 0.43071598413393986,
|
| 498 |
+
"learning_rate": 3e-06,
|
| 499 |
+
"loss": 0.4677,
|
| 500 |
+
"masked_tokens": 109.5875,
|
| 501 |
+
"mean_t": 0.5165087037021294,
|
| 502 |
+
"step": 190,
|
| 503 |
+
"student_masked_tokens": 109.5875
|
| 504 |
+
},
|
| 505 |
+
{
|
| 506 |
+
"avg_mask_ratio": 0.4665210062637925,
|
| 507 |
+
"avg_response_length": 197.8,
|
| 508 |
+
"avg_student_mask_ratio": 0.4665210062637925,
|
| 509 |
+
"batch_ainp_frac": 0.0,
|
| 510 |
+
"batch_inp_frac": 1.0,
|
| 511 |
+
"batch_inp_oh_frac": 0.0,
|
| 512 |
+
"batch_inp_par_frac": 0.0,
|
| 513 |
+
"batch_inp_par_par_frac": 0.0,
|
| 514 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 515 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 516 |
+
"batch_rl_frac": 0.0,
|
| 517 |
+
"batch_sft_frac": 0.0,
|
| 518 |
+
"batch_soft_sft_frac": 0.0,
|
| 519 |
+
"batch_tf_frac": 0.0,
|
| 520 |
+
"ce_loss": 0.0,
|
| 521 |
+
"epoch": 0.4266666666666667,
|
| 522 |
+
"grad_norm": 0.2421875,
|
| 523 |
+
"kd_loss": 0.35780135999341384,
|
| 524 |
+
"learning_rate": 3e-06,
|
| 525 |
+
"loss": 0.3577,
|
| 526 |
+
"masked_tokens": 97.0125,
|
| 527 |
+
"mean_t": 0.5073627714533359,
|
| 528 |
+
"step": 200,
|
| 529 |
+
"student_masked_tokens": 97.0125
|
| 530 |
+
},
|
| 531 |
+
{
|
| 532 |
+
"avg_mask_ratio": 0.45183838629163803,
|
| 533 |
+
"avg_response_length": 213.7875,
|
| 534 |
+
"avg_student_mask_ratio": 0.45183838629163803,
|
| 535 |
+
"batch_ainp_frac": 0.0,
|
| 536 |
+
"batch_inp_frac": 1.0,
|
| 537 |
+
"batch_inp_oh_frac": 0.0,
|
| 538 |
+
"batch_inp_par_frac": 0.0,
|
| 539 |
+
"batch_inp_par_par_frac": 0.0,
|
| 540 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 541 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 542 |
+
"batch_rl_frac": 0.0,
|
| 543 |
+
"batch_sft_frac": 0.0,
|
| 544 |
+
"batch_soft_sft_frac": 0.0,
|
| 545 |
+
"batch_tf_frac": 0.0,
|
| 546 |
+
"ce_loss": 0.0,
|
| 547 |
+
"epoch": 0.448,
|
| 548 |
+
"grad_norm": 0.2138671875,
|
| 549 |
+
"kd_loss": 0.38703344948080487,
|
| 550 |
+
"learning_rate": 3e-06,
|
| 551 |
+
"loss": 0.3627,
|
| 552 |
+
"masked_tokens": 96.0125,
|
| 553 |
+
"mean_t": 0.48430291628465055,
|
| 554 |
+
"step": 210,
|
| 555 |
+
"student_masked_tokens": 96.0125
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"avg_mask_ratio": 0.5396152360364794,
|
| 559 |
+
"avg_response_length": 236.5125,
|
| 560 |
+
"avg_student_mask_ratio": 0.5396152360364794,
|
| 561 |
+
"batch_ainp_frac": 0.0,
|
| 562 |
+
"batch_inp_frac": 1.0,
|
| 563 |
+
"batch_inp_oh_frac": 0.0,
|
| 564 |
+
"batch_inp_par_frac": 0.0,
|
| 565 |
+
"batch_inp_par_par_frac": 0.0,
|
| 566 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 567 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 568 |
+
"batch_rl_frac": 0.0,
|
| 569 |
+
"batch_sft_frac": 0.0,
|
| 570 |
+
"batch_soft_sft_frac": 0.0,
|
| 571 |
+
"batch_tf_frac": 0.0,
|
| 572 |
+
"ce_loss": 0.0,
|
| 573 |
+
"epoch": 0.4693333333333333,
|
| 574 |
+
"grad_norm": 0.3203125,
|
| 575 |
+
"kd_loss": 0.4665673110896023,
|
| 576 |
+
"learning_rate": 3e-06,
|
| 577 |
+
"loss": 0.5258,
|
| 578 |
+
"masked_tokens": 125.35,
|
| 579 |
+
"mean_t": 0.570199209311977,
|
| 580 |
+
"step": 220,
|
| 581 |
+
"student_masked_tokens": 125.35
|
| 582 |
+
},
|
| 583 |
+
{
|
| 584 |
+
"avg_mask_ratio": 0.4403991688624956,
|
| 585 |
+
"avg_response_length": 252.7,
|
| 586 |
+
"avg_student_mask_ratio": 0.4403991688624956,
|
| 587 |
+
"batch_ainp_frac": 0.0,
|
| 588 |
+
"batch_inp_frac": 1.0,
|
| 589 |
+
"batch_inp_oh_frac": 0.0,
|
| 590 |
+
"batch_inp_par_frac": 0.0,
|
| 591 |
+
"batch_inp_par_par_frac": 0.0,
|
| 592 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 593 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 594 |
+
"batch_rl_frac": 0.0,
|
| 595 |
+
"batch_sft_frac": 0.0,
|
| 596 |
+
"batch_soft_sft_frac": 0.0,
|
| 597 |
+
"batch_tf_frac": 0.0,
|
| 598 |
+
"ce_loss": 0.0,
|
| 599 |
+
"epoch": 0.49066666666666664,
|
| 600 |
+
"grad_norm": 0.220703125,
|
| 601 |
+
"kd_loss": 0.4347595912117505,
|
| 602 |
+
"learning_rate": 3e-06,
|
| 603 |
+
"loss": 0.4555,
|
| 604 |
+
"masked_tokens": 107.325,
|
| 605 |
+
"mean_t": 0.46891279935371133,
|
| 606 |
+
"step": 230,
|
| 607 |
+
"student_masked_tokens": 107.325
|
| 608 |
+
},
|
| 609 |
+
{
|
| 610 |
+
"avg_mask_ratio": 0.47419991258066146,
|
| 611 |
+
"avg_response_length": 212.85,
|
| 612 |
+
"avg_student_mask_ratio": 0.47419991258066146,
|
| 613 |
+
"batch_ainp_frac": 0.0,
|
| 614 |
+
"batch_inp_frac": 1.0,
|
| 615 |
+
"batch_inp_oh_frac": 0.0,
|
| 616 |
+
"batch_inp_par_frac": 0.0,
|
| 617 |
+
"batch_inp_par_par_frac": 0.0,
|
| 618 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 619 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 620 |
+
"batch_rl_frac": 0.0,
|
| 621 |
+
"batch_sft_frac": 0.0,
|
| 622 |
+
"batch_soft_sft_frac": 0.0,
|
| 623 |
+
"batch_tf_frac": 0.0,
|
| 624 |
+
"ce_loss": 0.0,
|
| 625 |
+
"epoch": 0.512,
|
| 626 |
+
"grad_norm": 0.1044921875,
|
| 627 |
+
"kd_loss": 0.5054432464686386,
|
| 628 |
+
"learning_rate": 3e-06,
|
| 629 |
+
"loss": 0.5307,
|
| 630 |
+
"masked_tokens": 106.9,
|
| 631 |
+
"mean_t": 0.502228345896583,
|
| 632 |
+
"step": 240,
|
| 633 |
+
"student_masked_tokens": 106.9
|
| 634 |
+
},
|
| 635 |
+
{
|
| 636 |
+
"avg_mask_ratio": 0.4464209079160355,
|
| 637 |
+
"avg_response_length": 243.475,
|
| 638 |
+
"avg_student_mask_ratio": 0.4464209079160355,
|
| 639 |
+
"batch_ainp_frac": 0.0,
|
| 640 |
+
"batch_inp_frac": 1.0,
|
| 641 |
+
"batch_inp_oh_frac": 0.0,
|
| 642 |
+
"batch_inp_par_frac": 0.0,
|
| 643 |
+
"batch_inp_par_par_frac": 0.0,
|
| 644 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 645 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 646 |
+
"batch_rl_frac": 0.0,
|
| 647 |
+
"batch_sft_frac": 0.0,
|
| 648 |
+
"batch_soft_sft_frac": 0.0,
|
| 649 |
+
"batch_tf_frac": 0.0,
|
| 650 |
+
"ce_loss": 0.0,
|
| 651 |
+
"epoch": 0.5333333333333333,
|
| 652 |
+
"grad_norm": 0.1025390625,
|
| 653 |
+
"kd_loss": 0.41803101517930374,
|
| 654 |
+
"learning_rate": 3e-06,
|
| 655 |
+
"loss": 0.3942,
|
| 656 |
+
"masked_tokens": 112.7375,
|
| 657 |
+
"mean_t": 0.4733429416548461,
|
| 658 |
+
"step": 250,
|
| 659 |
+
"student_masked_tokens": 112.7375
|
| 660 |
+
},
|
| 661 |
+
{
|
| 662 |
+
"avg_mask_ratio": 0.4520751796895638,
|
| 663 |
+
"avg_response_length": 245.55,
|
| 664 |
+
"avg_student_mask_ratio": 0.4520751796895638,
|
| 665 |
+
"batch_ainp_frac": 0.0,
|
| 666 |
+
"batch_inp_frac": 1.0,
|
| 667 |
+
"batch_inp_oh_frac": 0.0,
|
| 668 |
+
"batch_inp_par_frac": 0.0,
|
| 669 |
+
"batch_inp_par_par_frac": 0.0,
|
| 670 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 671 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 672 |
+
"batch_rl_frac": 0.0,
|
| 673 |
+
"batch_sft_frac": 0.0,
|
| 674 |
+
"batch_soft_sft_frac": 0.0,
|
| 675 |
+
"batch_tf_frac": 0.0,
|
| 676 |
+
"ce_loss": 0.0,
|
| 677 |
+
"epoch": 0.5546666666666666,
|
| 678 |
+
"grad_norm": 0.1806640625,
|
| 679 |
+
"kd_loss": 0.3202606088300026,
|
| 680 |
+
"learning_rate": 3e-06,
|
| 681 |
+
"loss": 0.3513,
|
| 682 |
+
"masked_tokens": 109.6375,
|
| 683 |
+
"mean_t": 0.4843149524240289,
|
| 684 |
+
"step": 260,
|
| 685 |
+
"student_masked_tokens": 109.6375
|
| 686 |
+
},
|
| 687 |
+
{
|
| 688 |
+
"avg_mask_ratio": 0.5305180630879477,
|
| 689 |
+
"avg_response_length": 224.45,
|
| 690 |
+
"avg_student_mask_ratio": 0.5305180630879477,
|
| 691 |
+
"batch_ainp_frac": 0.0,
|
| 692 |
+
"batch_inp_frac": 1.0,
|
| 693 |
+
"batch_inp_oh_frac": 0.0,
|
| 694 |
+
"batch_inp_par_frac": 0.0,
|
| 695 |
+
"batch_inp_par_par_frac": 0.0,
|
| 696 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 697 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 698 |
+
"batch_rl_frac": 0.0,
|
| 699 |
+
"batch_sft_frac": 0.0,
|
| 700 |
+
"batch_soft_sft_frac": 0.0,
|
| 701 |
+
"batch_tf_frac": 0.0,
|
| 702 |
+
"ce_loss": 0.0,
|
| 703 |
+
"epoch": 0.576,
|
| 704 |
+
"grad_norm": 0.28515625,
|
| 705 |
+
"kd_loss": 0.5523448550798549,
|
| 706 |
+
"learning_rate": 3e-06,
|
| 707 |
+
"loss": 0.5602,
|
| 708 |
+
"masked_tokens": 120.475,
|
| 709 |
+
"mean_t": 0.5643589949700981,
|
| 710 |
+
"step": 270,
|
| 711 |
+
"student_masked_tokens": 120.475
|
| 712 |
+
},
|
| 713 |
+
{
|
| 714 |
+
"avg_mask_ratio": 0.46451686368091033,
|
| 715 |
+
"avg_response_length": 254.825,
|
| 716 |
+
"avg_student_mask_ratio": 0.46451686368091033,
|
| 717 |
+
"batch_ainp_frac": 0.0,
|
| 718 |
+
"batch_inp_frac": 1.0,
|
| 719 |
+
"batch_inp_oh_frac": 0.0,
|
| 720 |
+
"batch_inp_par_frac": 0.0,
|
| 721 |
+
"batch_inp_par_par_frac": 0.0,
|
| 722 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 723 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 724 |
+
"batch_rl_frac": 0.0,
|
| 725 |
+
"batch_sft_frac": 0.0,
|
| 726 |
+
"batch_soft_sft_frac": 0.0,
|
| 727 |
+
"batch_tf_frac": 0.0,
|
| 728 |
+
"ce_loss": 0.0,
|
| 729 |
+
"epoch": 0.5973333333333334,
|
| 730 |
+
"grad_norm": 0.2451171875,
|
| 731 |
+
"kd_loss": 0.3936164884111079,
|
| 732 |
+
"learning_rate": 3e-06,
|
| 733 |
+
"loss": 0.4356,
|
| 734 |
+
"masked_tokens": 129.225,
|
| 735 |
+
"mean_t": 0.47818811538163575,
|
| 736 |
+
"step": 280,
|
| 737 |
+
"student_masked_tokens": 129.225
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"avg_mask_ratio": 0.389662017847877,
|
| 741 |
+
"avg_response_length": 245.9125,
|
| 742 |
+
"avg_student_mask_ratio": 0.389662017847877,
|
| 743 |
+
"batch_ainp_frac": 0.0,
|
| 744 |
+
"batch_inp_frac": 1.0,
|
| 745 |
+
"batch_inp_oh_frac": 0.0,
|
| 746 |
+
"batch_inp_par_frac": 0.0,
|
| 747 |
+
"batch_inp_par_par_frac": 0.0,
|
| 748 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 749 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 750 |
+
"batch_rl_frac": 0.0,
|
| 751 |
+
"batch_sft_frac": 0.0,
|
| 752 |
+
"batch_soft_sft_frac": 0.0,
|
| 753 |
+
"batch_tf_frac": 0.0,
|
| 754 |
+
"ce_loss": 0.0,
|
| 755 |
+
"epoch": 0.6186666666666667,
|
| 756 |
+
"grad_norm": 0.10888671875,
|
| 757 |
+
"kd_loss": 0.27756736385699055,
|
| 758 |
+
"learning_rate": 3e-06,
|
| 759 |
+
"loss": 0.3305,
|
| 760 |
+
"masked_tokens": 99.8625,
|
| 761 |
+
"mean_t": 0.4088635521940887,
|
| 762 |
+
"step": 290,
|
| 763 |
+
"student_masked_tokens": 99.8625
|
| 764 |
+
},
|
| 765 |
+
{
|
| 766 |
+
"avg_mask_ratio": 0.44417000194080175,
|
| 767 |
+
"avg_response_length": 217.0375,
|
| 768 |
+
"avg_student_mask_ratio": 0.44417000194080175,
|
| 769 |
+
"batch_ainp_frac": 0.0,
|
| 770 |
+
"batch_inp_frac": 1.0,
|
| 771 |
+
"batch_inp_oh_frac": 0.0,
|
| 772 |
+
"batch_inp_par_frac": 0.0,
|
| 773 |
+
"batch_inp_par_par_frac": 0.0,
|
| 774 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 775 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 776 |
+
"batch_rl_frac": 0.0,
|
| 777 |
+
"batch_sft_frac": 0.0,
|
| 778 |
+
"batch_soft_sft_frac": 0.0,
|
| 779 |
+
"batch_tf_frac": 0.0,
|
| 780 |
+
"ce_loss": 0.0,
|
| 781 |
+
"epoch": 0.64,
|
| 782 |
+
"grad_norm": 0.16015625,
|
| 783 |
+
"kd_loss": 0.31748922848000805,
|
| 784 |
+
"learning_rate": 3e-06,
|
| 785 |
+
"loss": 0.3705,
|
| 786 |
+
"masked_tokens": 104.625,
|
| 787 |
+
"mean_t": 0.47984200695063917,
|
| 788 |
+
"step": 300,
|
| 789 |
+
"student_masked_tokens": 104.625
|
| 790 |
+
},
|
| 791 |
+
{
|
| 792 |
+
"avg_mask_ratio": 0.4706685543409549,
|
| 793 |
+
"avg_response_length": 175.45,
|
| 794 |
+
"avg_student_mask_ratio": 0.4706685543409549,
|
| 795 |
+
"batch_ainp_frac": 0.0,
|
| 796 |
+
"batch_inp_frac": 1.0,
|
| 797 |
+
"batch_inp_oh_frac": 0.0,
|
| 798 |
+
"batch_inp_par_frac": 0.0,
|
| 799 |
+
"batch_inp_par_par_frac": 0.0,
|
| 800 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 801 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 802 |
+
"batch_rl_frac": 0.0,
|
| 803 |
+
"batch_sft_frac": 0.0,
|
| 804 |
+
"batch_soft_sft_frac": 0.0,
|
| 805 |
+
"batch_tf_frac": 0.0,
|
| 806 |
+
"ce_loss": 0.0,
|
| 807 |
+
"epoch": 0.6613333333333333,
|
| 808 |
+
"grad_norm": 0.263671875,
|
| 809 |
+
"kd_loss": 0.5043777307220239,
|
| 810 |
+
"learning_rate": 3e-06,
|
| 811 |
+
"loss": 0.4515,
|
| 812 |
+
"masked_tokens": 84.875,
|
| 813 |
+
"mean_t": 0.5026606284547597,
|
| 814 |
+
"step": 310,
|
| 815 |
+
"student_masked_tokens": 84.875
|
| 816 |
+
},
|
| 817 |
+
{
|
| 818 |
+
"avg_mask_ratio": 0.4974605386145413,
|
| 819 |
+
"avg_response_length": 234.7875,
|
| 820 |
+
"avg_student_mask_ratio": 0.4974605386145413,
|
| 821 |
+
"batch_ainp_frac": 0.0,
|
| 822 |
+
"batch_inp_frac": 1.0,
|
| 823 |
+
"batch_inp_oh_frac": 0.0,
|
| 824 |
+
"batch_inp_par_frac": 0.0,
|
| 825 |
+
"batch_inp_par_par_frac": 0.0,
|
| 826 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 827 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 828 |
+
"batch_rl_frac": 0.0,
|
| 829 |
+
"batch_sft_frac": 0.0,
|
| 830 |
+
"batch_soft_sft_frac": 0.0,
|
| 831 |
+
"batch_tf_frac": 0.0,
|
| 832 |
+
"ce_loss": 0.0,
|
| 833 |
+
"epoch": 0.6826666666666666,
|
| 834 |
+
"grad_norm": 0.287109375,
|
| 835 |
+
"kd_loss": 0.39651421944799947,
|
| 836 |
+
"learning_rate": 3e-06,
|
| 837 |
+
"loss": 0.4956,
|
| 838 |
+
"masked_tokens": 119.6,
|
| 839 |
+
"mean_t": 0.5293499688967132,
|
| 840 |
+
"step": 320,
|
| 841 |
+
"student_masked_tokens": 119.6
|
| 842 |
+
},
|
| 843 |
+
{
|
| 844 |
+
"avg_mask_ratio": 0.5112370474264025,
|
| 845 |
+
"avg_response_length": 236.0625,
|
| 846 |
+
"avg_student_mask_ratio": 0.5112370474264025,
|
| 847 |
+
"batch_ainp_frac": 0.0,
|
| 848 |
+
"batch_inp_frac": 1.0,
|
| 849 |
+
"batch_inp_oh_frac": 0.0,
|
| 850 |
+
"batch_inp_par_frac": 0.0,
|
| 851 |
+
"batch_inp_par_par_frac": 0.0,
|
| 852 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 853 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 854 |
+
"batch_rl_frac": 0.0,
|
| 855 |
+
"batch_sft_frac": 0.0,
|
| 856 |
+
"batch_soft_sft_frac": 0.0,
|
| 857 |
+
"batch_tf_frac": 0.0,
|
| 858 |
+
"ce_loss": 0.0,
|
| 859 |
+
"epoch": 0.704,
|
| 860 |
+
"grad_norm": 0.1982421875,
|
| 861 |
+
"kd_loss": 0.4267785593967403,
|
| 862 |
+
"learning_rate": 3e-06,
|
| 863 |
+
"loss": 0.4821,
|
| 864 |
+
"masked_tokens": 129.425,
|
| 865 |
+
"mean_t": 0.5426030711154454,
|
| 866 |
+
"step": 330,
|
| 867 |
+
"student_masked_tokens": 129.425
|
| 868 |
+
},
|
| 869 |
+
{
|
| 870 |
+
"avg_mask_ratio": 0.44370225080056114,
|
| 871 |
+
"avg_response_length": 241.4875,
|
| 872 |
+
"avg_student_mask_ratio": 0.44370225080056114,
|
| 873 |
+
"batch_ainp_frac": 0.0,
|
| 874 |
+
"batch_inp_frac": 1.0,
|
| 875 |
+
"batch_inp_oh_frac": 0.0,
|
| 876 |
+
"batch_inp_par_frac": 0.0,
|
| 877 |
+
"batch_inp_par_par_frac": 0.0,
|
| 878 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 879 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 880 |
+
"batch_rl_frac": 0.0,
|
| 881 |
+
"batch_sft_frac": 0.0,
|
| 882 |
+
"batch_soft_sft_frac": 0.0,
|
| 883 |
+
"batch_tf_frac": 0.0,
|
| 884 |
+
"ce_loss": 0.0,
|
| 885 |
+
"epoch": 0.7253333333333334,
|
| 886 |
+
"grad_norm": 0.306640625,
|
| 887 |
+
"kd_loss": 0.45950967887674776,
|
| 888 |
+
"learning_rate": 3e-06,
|
| 889 |
+
"loss": 0.4426,
|
| 890 |
+
"masked_tokens": 108.775,
|
| 891 |
+
"mean_t": 0.47635243807453664,
|
| 892 |
+
"step": 340,
|
| 893 |
+
"student_masked_tokens": 108.775
|
| 894 |
+
},
|
| 895 |
+
{
|
| 896 |
+
"avg_mask_ratio": 0.49959173843380994,
|
| 897 |
+
"avg_response_length": 235.6375,
|
| 898 |
+
"avg_student_mask_ratio": 0.49959173843380994,
|
| 899 |
+
"batch_ainp_frac": 0.0,
|
| 900 |
+
"batch_inp_frac": 1.0,
|
| 901 |
+
"batch_inp_oh_frac": 0.0,
|
| 902 |
+
"batch_inp_par_frac": 0.0,
|
| 903 |
+
"batch_inp_par_par_frac": 0.0,
|
| 904 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 905 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 906 |
+
"batch_rl_frac": 0.0,
|
| 907 |
+
"batch_sft_frac": 0.0,
|
| 908 |
+
"batch_soft_sft_frac": 0.0,
|
| 909 |
+
"batch_tf_frac": 0.0,
|
| 910 |
+
"ce_loss": 0.0,
|
| 911 |
+
"epoch": 0.7466666666666667,
|
| 912 |
+
"grad_norm": 0.46875,
|
| 913 |
+
"kd_loss": 0.5008536678237376,
|
| 914 |
+
"learning_rate": 3e-06,
|
| 915 |
+
"loss": 0.5298,
|
| 916 |
+
"masked_tokens": 125.625,
|
| 917 |
+
"mean_t": 0.5268881446914747,
|
| 918 |
+
"step": 350,
|
| 919 |
+
"student_masked_tokens": 125.625
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"avg_mask_ratio": 0.4744729608530179,
|
| 923 |
+
"avg_response_length": 246.1625,
|
| 924 |
+
"avg_student_mask_ratio": 0.4744729608530179,
|
| 925 |
+
"batch_ainp_frac": 0.0,
|
| 926 |
+
"batch_inp_frac": 1.0,
|
| 927 |
+
"batch_inp_oh_frac": 0.0,
|
| 928 |
+
"batch_inp_par_frac": 0.0,
|
| 929 |
+
"batch_inp_par_par_frac": 0.0,
|
| 930 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 931 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 932 |
+
"batch_rl_frac": 0.0,
|
| 933 |
+
"batch_sft_frac": 0.0,
|
| 934 |
+
"batch_soft_sft_frac": 0.0,
|
| 935 |
+
"batch_tf_frac": 0.0,
|
| 936 |
+
"ce_loss": 0.0,
|
| 937 |
+
"epoch": 0.768,
|
| 938 |
+
"grad_norm": 0.2099609375,
|
| 939 |
+
"kd_loss": 0.5271339516211748,
|
| 940 |
+
"learning_rate": 3e-06,
|
| 941 |
+
"loss": 0.4601,
|
| 942 |
+
"masked_tokens": 116.875,
|
| 943 |
+
"mean_t": 0.5040419134311378,
|
| 944 |
+
"step": 360,
|
| 945 |
+
"student_masked_tokens": 116.875
|
| 946 |
+
},
|
| 947 |
+
{
|
| 948 |
+
"avg_mask_ratio": 0.4738045462174341,
|
| 949 |
+
"avg_response_length": 257.575,
|
| 950 |
+
"avg_student_mask_ratio": 0.4738045462174341,
|
| 951 |
+
"batch_ainp_frac": 0.0,
|
| 952 |
+
"batch_inp_frac": 1.0,
|
| 953 |
+
"batch_inp_oh_frac": 0.0,
|
| 954 |
+
"batch_inp_par_frac": 0.0,
|
| 955 |
+
"batch_inp_par_par_frac": 0.0,
|
| 956 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 957 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 958 |
+
"batch_rl_frac": 0.0,
|
| 959 |
+
"batch_sft_frac": 0.0,
|
| 960 |
+
"batch_soft_sft_frac": 0.0,
|
| 961 |
+
"batch_tf_frac": 0.0,
|
| 962 |
+
"ce_loss": 0.0,
|
| 963 |
+
"epoch": 0.7893333333333333,
|
| 964 |
+
"grad_norm": 0.166015625,
|
| 965 |
+
"kd_loss": 0.6038310030206958,
|
| 966 |
+
"learning_rate": 3e-06,
|
| 967 |
+
"loss": 0.4979,
|
| 968 |
+
"masked_tokens": 127.4625,
|
| 969 |
+
"mean_t": 0.5127181728370488,
|
| 970 |
+
"step": 370,
|
| 971 |
+
"student_masked_tokens": 127.4625
|
| 972 |
+
},
|
| 973 |
+
{
|
| 974 |
+
"avg_mask_ratio": 0.4512475330149755,
|
| 975 |
+
"avg_response_length": 209.8,
|
| 976 |
+
"avg_student_mask_ratio": 0.4512475330149755,
|
| 977 |
+
"batch_ainp_frac": 0.0,
|
| 978 |
+
"batch_inp_frac": 1.0,
|
| 979 |
+
"batch_inp_oh_frac": 0.0,
|
| 980 |
+
"batch_inp_par_frac": 0.0,
|
| 981 |
+
"batch_inp_par_par_frac": 0.0,
|
| 982 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 983 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 984 |
+
"batch_rl_frac": 0.0,
|
| 985 |
+
"batch_sft_frac": 0.0,
|
| 986 |
+
"batch_soft_sft_frac": 0.0,
|
| 987 |
+
"batch_tf_frac": 0.0,
|
| 988 |
+
"ce_loss": 0.0,
|
| 989 |
+
"epoch": 0.8106666666666666,
|
| 990 |
+
"grad_norm": 0.20703125,
|
| 991 |
+
"kd_loss": 0.4014330595062404,
|
| 992 |
+
"learning_rate": 3e-06,
|
| 993 |
+
"loss": 0.3648,
|
| 994 |
+
"masked_tokens": 100.8375,
|
| 995 |
+
"mean_t": 0.4825185665744357,
|
| 996 |
+
"step": 380,
|
| 997 |
+
"student_masked_tokens": 100.8375
|
| 998 |
+
},
|
| 999 |
+
{
|
| 1000 |
+
"avg_mask_ratio": 0.4752940105390735,
|
| 1001 |
+
"avg_response_length": 219.5625,
|
| 1002 |
+
"avg_student_mask_ratio": 0.4752940105390735,
|
| 1003 |
+
"batch_ainp_frac": 0.0,
|
| 1004 |
+
"batch_inp_frac": 1.0,
|
| 1005 |
+
"batch_inp_oh_frac": 0.0,
|
| 1006 |
+
"batch_inp_par_frac": 0.0,
|
| 1007 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1008 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1009 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1010 |
+
"batch_rl_frac": 0.0,
|
| 1011 |
+
"batch_sft_frac": 0.0,
|
| 1012 |
+
"batch_soft_sft_frac": 0.0,
|
| 1013 |
+
"batch_tf_frac": 0.0,
|
| 1014 |
+
"ce_loss": 0.0,
|
| 1015 |
+
"epoch": 0.832,
|
| 1016 |
+
"grad_norm": 0.1484375,
|
| 1017 |
+
"kd_loss": 0.4614308591259089,
|
| 1018 |
+
"learning_rate": 3e-06,
|
| 1019 |
+
"loss": 0.465,
|
| 1020 |
+
"masked_tokens": 112.375,
|
| 1021 |
+
"mean_t": 0.5053101469413377,
|
| 1022 |
+
"step": 390,
|
| 1023 |
+
"student_masked_tokens": 112.375
|
| 1024 |
+
},
|
| 1025 |
+
{
|
| 1026 |
+
"avg_mask_ratio": 0.47461870914557946,
|
| 1027 |
+
"avg_response_length": 242.6125,
|
| 1028 |
+
"avg_student_mask_ratio": 0.47461870914557946,
|
| 1029 |
+
"batch_ainp_frac": 0.0,
|
| 1030 |
+
"batch_inp_frac": 1.0,
|
| 1031 |
+
"batch_inp_oh_frac": 0.0,
|
| 1032 |
+
"batch_inp_par_frac": 0.0,
|
| 1033 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1034 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1035 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1036 |
+
"batch_rl_frac": 0.0,
|
| 1037 |
+
"batch_sft_frac": 0.0,
|
| 1038 |
+
"batch_soft_sft_frac": 0.0,
|
| 1039 |
+
"batch_tf_frac": 0.0,
|
| 1040 |
+
"ce_loss": 0.0,
|
| 1041 |
+
"epoch": 0.8533333333333334,
|
| 1042 |
+
"grad_norm": 0.3671875,
|
| 1043 |
+
"kd_loss": 0.5285846941962745,
|
| 1044 |
+
"learning_rate": 3e-06,
|
| 1045 |
+
"loss": 0.4838,
|
| 1046 |
+
"masked_tokens": 120.4125,
|
| 1047 |
+
"mean_t": 0.5052250675857067,
|
| 1048 |
+
"step": 400,
|
| 1049 |
+
"student_masked_tokens": 120.4125
|
| 1050 |
+
},
|
| 1051 |
+
{
|
| 1052 |
+
"avg_mask_ratio": 0.48321815438685006,
|
| 1053 |
+
"avg_response_length": 228.15,
|
| 1054 |
+
"avg_student_mask_ratio": 0.48321815438685006,
|
| 1055 |
+
"batch_ainp_frac": 0.0,
|
| 1056 |
+
"batch_inp_frac": 1.0,
|
| 1057 |
+
"batch_inp_oh_frac": 0.0,
|
| 1058 |
+
"batch_inp_par_frac": 0.0,
|
| 1059 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1060 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1061 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1062 |
+
"batch_rl_frac": 0.0,
|
| 1063 |
+
"batch_sft_frac": 0.0,
|
| 1064 |
+
"batch_soft_sft_frac": 0.0,
|
| 1065 |
+
"batch_tf_frac": 0.0,
|
| 1066 |
+
"ce_loss": 0.0,
|
| 1067 |
+
"epoch": 0.8746666666666667,
|
| 1068 |
+
"grad_norm": 0.2353515625,
|
| 1069 |
+
"kd_loss": 0.5044134976945088,
|
| 1070 |
+
"learning_rate": 3e-06,
|
| 1071 |
+
"loss": 0.5269,
|
| 1072 |
+
"masked_tokens": 119.0,
|
| 1073 |
+
"mean_t": 0.5050956419203431,
|
| 1074 |
+
"step": 410,
|
| 1075 |
+
"student_masked_tokens": 119.0
|
| 1076 |
+
},
|
| 1077 |
+
{
|
| 1078 |
+
"avg_mask_ratio": 0.4379329536575824,
|
| 1079 |
+
"avg_response_length": 220.225,
|
| 1080 |
+
"avg_student_mask_ratio": 0.4379329536575824,
|
| 1081 |
+
"batch_ainp_frac": 0.0,
|
| 1082 |
+
"batch_inp_frac": 1.0,
|
| 1083 |
+
"batch_inp_oh_frac": 0.0,
|
| 1084 |
+
"batch_inp_par_frac": 0.0,
|
| 1085 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1086 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1087 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1088 |
+
"batch_rl_frac": 0.0,
|
| 1089 |
+
"batch_sft_frac": 0.0,
|
| 1090 |
+
"batch_soft_sft_frac": 0.0,
|
| 1091 |
+
"batch_tf_frac": 0.0,
|
| 1092 |
+
"ce_loss": 0.0,
|
| 1093 |
+
"epoch": 0.896,
|
| 1094 |
+
"grad_norm": 0.1640625,
|
| 1095 |
+
"kd_loss": 0.27499000265762563,
|
| 1096 |
+
"learning_rate": 3e-06,
|
| 1097 |
+
"loss": 0.2998,
|
| 1098 |
+
"masked_tokens": 85.525,
|
| 1099 |
+
"mean_t": 0.4769687672611326,
|
| 1100 |
+
"step": 420,
|
| 1101 |
+
"student_masked_tokens": 85.525
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"avg_mask_ratio": 0.4674084897618741,
|
| 1105 |
+
"avg_response_length": 249.2125,
|
| 1106 |
+
"avg_student_mask_ratio": 0.4674084897618741,
|
| 1107 |
+
"batch_ainp_frac": 0.0,
|
| 1108 |
+
"batch_inp_frac": 1.0,
|
| 1109 |
+
"batch_inp_oh_frac": 0.0,
|
| 1110 |
+
"batch_inp_par_frac": 0.0,
|
| 1111 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1112 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1113 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1114 |
+
"batch_rl_frac": 0.0,
|
| 1115 |
+
"batch_sft_frac": 0.0,
|
| 1116 |
+
"batch_soft_sft_frac": 0.0,
|
| 1117 |
+
"batch_tf_frac": 0.0,
|
| 1118 |
+
"ce_loss": 0.0,
|
| 1119 |
+
"epoch": 0.9173333333333333,
|
| 1120 |
+
"grad_norm": 0.2255859375,
|
| 1121 |
+
"kd_loss": 0.4936492626361996,
|
| 1122 |
+
"learning_rate": 3e-06,
|
| 1123 |
+
"loss": 0.4862,
|
| 1124 |
+
"masked_tokens": 104.5625,
|
| 1125 |
+
"mean_t": 0.49262027950026094,
|
| 1126 |
+
"step": 430,
|
| 1127 |
+
"student_masked_tokens": 104.5625
|
| 1128 |
+
},
|
| 1129 |
+
{
|
| 1130 |
+
"avg_mask_ratio": 0.4415457699564286,
|
| 1131 |
+
"avg_response_length": 241.0875,
|
| 1132 |
+
"avg_student_mask_ratio": 0.4415457699564286,
|
| 1133 |
+
"batch_ainp_frac": 0.0,
|
| 1134 |
+
"batch_inp_frac": 1.0,
|
| 1135 |
+
"batch_inp_oh_frac": 0.0,
|
| 1136 |
+
"batch_inp_par_frac": 0.0,
|
| 1137 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1138 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1139 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1140 |
+
"batch_rl_frac": 0.0,
|
| 1141 |
+
"batch_sft_frac": 0.0,
|
| 1142 |
+
"batch_soft_sft_frac": 0.0,
|
| 1143 |
+
"batch_tf_frac": 0.0,
|
| 1144 |
+
"ce_loss": 0.0,
|
| 1145 |
+
"epoch": 0.9386666666666666,
|
| 1146 |
+
"grad_norm": 0.2265625,
|
| 1147 |
+
"kd_loss": 0.4476076794173423,
|
| 1148 |
+
"learning_rate": 3e-06,
|
| 1149 |
+
"loss": 0.4732,
|
| 1150 |
+
"masked_tokens": 113.0875,
|
| 1151 |
+
"mean_t": 0.47046207524836064,
|
| 1152 |
+
"step": 440,
|
| 1153 |
+
"student_masked_tokens": 113.0875
|
| 1154 |
+
},
|
| 1155 |
+
{
|
| 1156 |
+
"avg_mask_ratio": 0.42486972180195154,
|
| 1157 |
+
"avg_response_length": 231.9875,
|
| 1158 |
+
"avg_student_mask_ratio": 0.42486972180195154,
|
| 1159 |
+
"batch_ainp_frac": 0.0,
|
| 1160 |
+
"batch_inp_frac": 1.0,
|
| 1161 |
+
"batch_inp_oh_frac": 0.0,
|
| 1162 |
+
"batch_inp_par_frac": 0.0,
|
| 1163 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1164 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1165 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1166 |
+
"batch_rl_frac": 0.0,
|
| 1167 |
+
"batch_sft_frac": 0.0,
|
| 1168 |
+
"batch_soft_sft_frac": 0.0,
|
| 1169 |
+
"batch_tf_frac": 0.0,
|
| 1170 |
+
"ce_loss": 0.0,
|
| 1171 |
+
"epoch": 0.96,
|
| 1172 |
+
"grad_norm": 0.255859375,
|
| 1173 |
+
"kd_loss": 0.40184078344118745,
|
| 1174 |
+
"learning_rate": 3e-06,
|
| 1175 |
+
"loss": 0.3814,
|
| 1176 |
+
"masked_tokens": 103.4,
|
| 1177 |
+
"mean_t": 0.45781184462830427,
|
| 1178 |
+
"step": 450,
|
| 1179 |
+
"student_masked_tokens": 103.4
|
| 1180 |
+
},
|
| 1181 |
+
{
|
| 1182 |
+
"avg_mask_ratio": 0.47578654896933587,
|
| 1183 |
+
"avg_response_length": 214.6125,
|
| 1184 |
+
"avg_student_mask_ratio": 0.47578654896933587,
|
| 1185 |
+
"batch_ainp_frac": 0.0,
|
| 1186 |
+
"batch_inp_frac": 1.0,
|
| 1187 |
+
"batch_inp_oh_frac": 0.0,
|
| 1188 |
+
"batch_inp_par_frac": 0.0,
|
| 1189 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1190 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1191 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1192 |
+
"batch_rl_frac": 0.0,
|
| 1193 |
+
"batch_sft_frac": 0.0,
|
| 1194 |
+
"batch_soft_sft_frac": 0.0,
|
| 1195 |
+
"batch_tf_frac": 0.0,
|
| 1196 |
+
"ce_loss": 0.0,
|
| 1197 |
+
"epoch": 0.9813333333333333,
|
| 1198 |
+
"grad_norm": 0.15625,
|
| 1199 |
+
"kd_loss": 0.43999221579811093,
|
| 1200 |
+
"learning_rate": 3e-06,
|
| 1201 |
+
"loss": 0.45,
|
| 1202 |
+
"masked_tokens": 105.3125,
|
| 1203 |
+
"mean_t": 0.5075790266972036,
|
| 1204 |
+
"step": 460,
|
| 1205 |
+
"student_masked_tokens": 105.3125
|
| 1206 |
+
},
|
| 1207 |
+
{
|
| 1208 |
+
"avg_mask_ratio": 0.4782901787132557,
|
| 1209 |
+
"avg_response_length": 224.0952380952381,
|
| 1210 |
+
"avg_student_mask_ratio": 0.4782901787132557,
|
| 1211 |
+
"batch_ainp_frac": 0.0,
|
| 1212 |
+
"batch_inp_frac": 1.0,
|
| 1213 |
+
"batch_inp_oh_frac": 0.0,
|
| 1214 |
+
"batch_inp_par_frac": 0.0,
|
| 1215 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1216 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1217 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1218 |
+
"batch_rl_frac": 0.0,
|
| 1219 |
+
"batch_sft_frac": 0.0,
|
| 1220 |
+
"batch_soft_sft_frac": 0.0,
|
| 1221 |
+
"batch_tf_frac": 0.0,
|
| 1222 |
+
"ce_loss": 0.0,
|
| 1223 |
+
"epoch": 1.0042666666666666,
|
| 1224 |
+
"grad_norm": 0.330078125,
|
| 1225 |
+
"kd_loss": 0.5121579658046268,
|
| 1226 |
+
"learning_rate": 3e-06,
|
| 1227 |
+
"loss": 0.5319,
|
| 1228 |
+
"masked_tokens": 107.23809523809524,
|
| 1229 |
+
"mean_t": 0.5031429776822084,
|
| 1230 |
+
"step": 470,
|
| 1231 |
+
"student_masked_tokens": 107.23809523809524
|
| 1232 |
+
},
|
| 1233 |
+
{
|
| 1234 |
+
"avg_mask_ratio": 0.47575968883465974,
|
| 1235 |
+
"avg_response_length": 249.4125,
|
| 1236 |
+
"avg_student_mask_ratio": 0.47575968883465974,
|
| 1237 |
+
"batch_ainp_frac": 0.0,
|
| 1238 |
+
"batch_inp_frac": 1.0,
|
| 1239 |
+
"batch_inp_oh_frac": 0.0,
|
| 1240 |
+
"batch_inp_par_frac": 0.0,
|
| 1241 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1242 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1243 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1244 |
+
"batch_rl_frac": 0.0,
|
| 1245 |
+
"batch_sft_frac": 0.0,
|
| 1246 |
+
"batch_soft_sft_frac": 0.0,
|
| 1247 |
+
"batch_tf_frac": 0.0,
|
| 1248 |
+
"ce_loss": 0.0,
|
| 1249 |
+
"epoch": 1.0256,
|
| 1250 |
+
"grad_norm": 0.220703125,
|
| 1251 |
+
"kd_loss": 0.5364690735516149,
|
| 1252 |
+
"learning_rate": 3e-06,
|
| 1253 |
+
"loss": 0.4899,
|
| 1254 |
+
"masked_tokens": 118.35,
|
| 1255 |
+
"mean_t": 0.504472183593316,
|
| 1256 |
+
"step": 480,
|
| 1257 |
+
"student_masked_tokens": 118.35
|
| 1258 |
+
},
|
| 1259 |
+
{
|
| 1260 |
+
"avg_mask_ratio": 0.4563717324635945,
|
| 1261 |
+
"avg_response_length": 232.0375,
|
| 1262 |
+
"avg_student_mask_ratio": 0.4563717324635945,
|
| 1263 |
+
"batch_ainp_frac": 0.0,
|
| 1264 |
+
"batch_inp_frac": 1.0,
|
| 1265 |
+
"batch_inp_oh_frac": 0.0,
|
| 1266 |
+
"batch_inp_par_frac": 0.0,
|
| 1267 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1268 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1269 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1270 |
+
"batch_rl_frac": 0.0,
|
| 1271 |
+
"batch_sft_frac": 0.0,
|
| 1272 |
+
"batch_soft_sft_frac": 0.0,
|
| 1273 |
+
"batch_tf_frac": 0.0,
|
| 1274 |
+
"ce_loss": 0.0,
|
| 1275 |
+
"epoch": 1.0469333333333333,
|
| 1276 |
+
"grad_norm": 0.39453125,
|
| 1277 |
+
"kd_loss": 0.38637587467336515,
|
| 1278 |
+
"learning_rate": 3e-06,
|
| 1279 |
+
"loss": 0.4045,
|
| 1280 |
+
"masked_tokens": 98.35,
|
| 1281 |
+
"mean_t": 0.4888980514719151,
|
| 1282 |
+
"step": 490,
|
| 1283 |
+
"student_masked_tokens": 98.35
|
| 1284 |
+
},
|
| 1285 |
+
{
|
| 1286 |
+
"avg_mask_ratio": 0.5079968665260821,
|
| 1287 |
+
"avg_response_length": 253.7875,
|
| 1288 |
+
"avg_student_mask_ratio": 0.5079968665260821,
|
| 1289 |
+
"batch_ainp_frac": 0.0,
|
| 1290 |
+
"batch_inp_frac": 1.0,
|
| 1291 |
+
"batch_inp_oh_frac": 0.0,
|
| 1292 |
+
"batch_inp_par_frac": 0.0,
|
| 1293 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1294 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1295 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1296 |
+
"batch_rl_frac": 0.0,
|
| 1297 |
+
"batch_sft_frac": 0.0,
|
| 1298 |
+
"batch_soft_sft_frac": 0.0,
|
| 1299 |
+
"batch_tf_frac": 0.0,
|
| 1300 |
+
"ce_loss": 0.0,
|
| 1301 |
+
"epoch": 1.0682666666666667,
|
| 1302 |
+
"grad_norm": 0.26171875,
|
| 1303 |
+
"kd_loss": 0.4493013881629054,
|
| 1304 |
+
"learning_rate": 3e-06,
|
| 1305 |
+
"loss": 0.4348,
|
| 1306 |
+
"masked_tokens": 128.225,
|
| 1307 |
+
"mean_t": 0.5469163245841628,
|
| 1308 |
+
"step": 500,
|
| 1309 |
+
"student_masked_tokens": 128.225
|
| 1310 |
+
},
|
| 1311 |
+
{
|
| 1312 |
+
"avg_mask_ratio": 0.5109448074479588,
|
| 1313 |
+
"avg_response_length": 254.2,
|
| 1314 |
+
"avg_student_mask_ratio": 0.5109448074479588,
|
| 1315 |
+
"batch_ainp_frac": 0.0,
|
| 1316 |
+
"batch_inp_frac": 1.0,
|
| 1317 |
+
"batch_inp_oh_frac": 0.0,
|
| 1318 |
+
"batch_inp_par_frac": 0.0,
|
| 1319 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1320 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1321 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1322 |
+
"batch_rl_frac": 0.0,
|
| 1323 |
+
"batch_sft_frac": 0.0,
|
| 1324 |
+
"batch_soft_sft_frac": 0.0,
|
| 1325 |
+
"batch_tf_frac": 0.0,
|
| 1326 |
+
"ce_loss": 0.0,
|
| 1327 |
+
"epoch": 1.0896,
|
| 1328 |
+
"grad_norm": 0.2255859375,
|
| 1329 |
+
"kd_loss": 0.5702315452758938,
|
| 1330 |
+
"learning_rate": 3e-06,
|
| 1331 |
+
"loss": 0.5036,
|
| 1332 |
+
"masked_tokens": 137.5875,
|
| 1333 |
+
"mean_t": 0.5275314710394013,
|
| 1334 |
+
"step": 510,
|
| 1335 |
+
"student_masked_tokens": 137.5875
|
| 1336 |
+
},
|
| 1337 |
+
{
|
| 1338 |
+
"avg_mask_ratio": 0.45396183808334173,
|
| 1339 |
+
"avg_response_length": 202.7625,
|
| 1340 |
+
"avg_student_mask_ratio": 0.45396183808334173,
|
| 1341 |
+
"batch_ainp_frac": 0.0,
|
| 1342 |
+
"batch_inp_frac": 1.0,
|
| 1343 |
+
"batch_inp_oh_frac": 0.0,
|
| 1344 |
+
"batch_inp_par_frac": 0.0,
|
| 1345 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1346 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1347 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1348 |
+
"batch_rl_frac": 0.0,
|
| 1349 |
+
"batch_sft_frac": 0.0,
|
| 1350 |
+
"batch_soft_sft_frac": 0.0,
|
| 1351 |
+
"batch_tf_frac": 0.0,
|
| 1352 |
+
"ce_loss": 0.0,
|
| 1353 |
+
"epoch": 1.1109333333333333,
|
| 1354 |
+
"grad_norm": 0.2421875,
|
| 1355 |
+
"kd_loss": 0.420669932495889,
|
| 1356 |
+
"learning_rate": 3e-06,
|
| 1357 |
+
"loss": 0.4283,
|
| 1358 |
+
"masked_tokens": 87.0625,
|
| 1359 |
+
"mean_t": 0.49193521235138177,
|
| 1360 |
+
"step": 520,
|
| 1361 |
+
"student_masked_tokens": 87.0625
|
| 1362 |
+
},
|
| 1363 |
+
{
|
| 1364 |
+
"avg_mask_ratio": 0.47015948037151245,
|
| 1365 |
+
"avg_response_length": 214.275,
|
| 1366 |
+
"avg_student_mask_ratio": 0.47015948037151245,
|
| 1367 |
+
"batch_ainp_frac": 0.0,
|
| 1368 |
+
"batch_inp_frac": 1.0,
|
| 1369 |
+
"batch_inp_oh_frac": 0.0,
|
| 1370 |
+
"batch_inp_par_frac": 0.0,
|
| 1371 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1372 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1373 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1374 |
+
"batch_rl_frac": 0.0,
|
| 1375 |
+
"batch_sft_frac": 0.0,
|
| 1376 |
+
"batch_soft_sft_frac": 0.0,
|
| 1377 |
+
"batch_tf_frac": 0.0,
|
| 1378 |
+
"ce_loss": 0.0,
|
| 1379 |
+
"epoch": 1.1322666666666668,
|
| 1380 |
+
"grad_norm": 0.31640625,
|
| 1381 |
+
"kd_loss": 0.4567407854818356,
|
| 1382 |
+
"learning_rate": 3e-06,
|
| 1383 |
+
"loss": 0.5135,
|
| 1384 |
+
"masked_tokens": 99.8625,
|
| 1385 |
+
"mean_t": 0.49621942077938,
|
| 1386 |
+
"step": 530,
|
| 1387 |
+
"student_masked_tokens": 99.8625
|
| 1388 |
+
},
|
| 1389 |
+
{
|
| 1390 |
+
"avg_mask_ratio": 0.4892866689246148,
|
| 1391 |
+
"avg_response_length": 231.3125,
|
| 1392 |
+
"avg_student_mask_ratio": 0.4892866689246148,
|
| 1393 |
+
"batch_ainp_frac": 0.0,
|
| 1394 |
+
"batch_inp_frac": 1.0,
|
| 1395 |
+
"batch_inp_oh_frac": 0.0,
|
| 1396 |
+
"batch_inp_par_frac": 0.0,
|
| 1397 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1398 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1399 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1400 |
+
"batch_rl_frac": 0.0,
|
| 1401 |
+
"batch_sft_frac": 0.0,
|
| 1402 |
+
"batch_soft_sft_frac": 0.0,
|
| 1403 |
+
"batch_tf_frac": 0.0,
|
| 1404 |
+
"ce_loss": 0.0,
|
| 1405 |
+
"epoch": 1.1536,
|
| 1406 |
+
"grad_norm": 0.3359375,
|
| 1407 |
+
"kd_loss": 0.5524838666212872,
|
| 1408 |
+
"learning_rate": 3e-06,
|
| 1409 |
+
"loss": 0.5048,
|
| 1410 |
+
"masked_tokens": 107.375,
|
| 1411 |
+
"mean_t": 0.5208023569080978,
|
| 1412 |
+
"step": 540,
|
| 1413 |
+
"student_masked_tokens": 107.375
|
| 1414 |
+
},
|
| 1415 |
+
{
|
| 1416 |
+
"avg_mask_ratio": 0.4541942774085328,
|
| 1417 |
+
"avg_response_length": 213.525,
|
| 1418 |
+
"avg_student_mask_ratio": 0.4541942774085328,
|
| 1419 |
+
"batch_ainp_frac": 0.0,
|
| 1420 |
+
"batch_inp_frac": 1.0,
|
| 1421 |
+
"batch_inp_oh_frac": 0.0,
|
| 1422 |
+
"batch_inp_par_frac": 0.0,
|
| 1423 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1424 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1425 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1426 |
+
"batch_rl_frac": 0.0,
|
| 1427 |
+
"batch_sft_frac": 0.0,
|
| 1428 |
+
"batch_soft_sft_frac": 0.0,
|
| 1429 |
+
"batch_tf_frac": 0.0,
|
| 1430 |
+
"ce_loss": 0.0,
|
| 1431 |
+
"epoch": 1.1749333333333334,
|
| 1432 |
+
"grad_norm": 0.1572265625,
|
| 1433 |
+
"kd_loss": 0.3688938313097424,
|
| 1434 |
+
"learning_rate": 3e-06,
|
| 1435 |
+
"loss": 0.3777,
|
| 1436 |
+
"masked_tokens": 97.8875,
|
| 1437 |
+
"mean_t": 0.4767197913257405,
|
| 1438 |
+
"step": 550,
|
| 1439 |
+
"student_masked_tokens": 97.8875
|
| 1440 |
+
},
|
| 1441 |
+
{
|
| 1442 |
+
"avg_mask_ratio": 0.39282396506750955,
|
| 1443 |
+
"avg_response_length": 231.4125,
|
| 1444 |
+
"avg_student_mask_ratio": 0.39282396506750955,
|
| 1445 |
+
"batch_ainp_frac": 0.0,
|
| 1446 |
+
"batch_inp_frac": 1.0,
|
| 1447 |
+
"batch_inp_oh_frac": 0.0,
|
| 1448 |
+
"batch_inp_par_frac": 0.0,
|
| 1449 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1450 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1451 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1452 |
+
"batch_rl_frac": 0.0,
|
| 1453 |
+
"batch_sft_frac": 0.0,
|
| 1454 |
+
"batch_soft_sft_frac": 0.0,
|
| 1455 |
+
"batch_tf_frac": 0.0,
|
| 1456 |
+
"ce_loss": 0.0,
|
| 1457 |
+
"epoch": 1.1962666666666666,
|
| 1458 |
+
"grad_norm": 0.359375,
|
| 1459 |
+
"kd_loss": 0.4847641052244853,
|
| 1460 |
+
"learning_rate": 3e-06,
|
| 1461 |
+
"loss": 0.4135,
|
| 1462 |
+
"masked_tokens": 99.2875,
|
| 1463 |
+
"mean_t": 0.4111072298779618,
|
| 1464 |
+
"step": 560,
|
| 1465 |
+
"student_masked_tokens": 99.2875
|
| 1466 |
+
},
|
| 1467 |
+
{
|
| 1468 |
+
"avg_mask_ratio": 0.4483634108910337,
|
| 1469 |
+
"avg_response_length": 230.1625,
|
| 1470 |
+
"avg_student_mask_ratio": 0.4483634108910337,
|
| 1471 |
+
"batch_ainp_frac": 0.0,
|
| 1472 |
+
"batch_inp_frac": 1.0,
|
| 1473 |
+
"batch_inp_oh_frac": 0.0,
|
| 1474 |
+
"batch_inp_par_frac": 0.0,
|
| 1475 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1476 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1477 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1478 |
+
"batch_rl_frac": 0.0,
|
| 1479 |
+
"batch_sft_frac": 0.0,
|
| 1480 |
+
"batch_soft_sft_frac": 0.0,
|
| 1481 |
+
"batch_tf_frac": 0.0,
|
| 1482 |
+
"ce_loss": 0.0,
|
| 1483 |
+
"epoch": 1.2176,
|
| 1484 |
+
"grad_norm": 0.2294921875,
|
| 1485 |
+
"kd_loss": 0.41131700558344164,
|
| 1486 |
+
"learning_rate": 3e-06,
|
| 1487 |
+
"loss": 0.4323,
|
| 1488 |
+
"masked_tokens": 108.9875,
|
| 1489 |
+
"mean_t": 0.48533305872697385,
|
| 1490 |
+
"step": 570,
|
| 1491 |
+
"student_masked_tokens": 108.9875
|
| 1492 |
+
},
|
| 1493 |
+
{
|
| 1494 |
+
"avg_mask_ratio": 0.465452536707744,
|
| 1495 |
+
"avg_response_length": 267.4375,
|
| 1496 |
+
"avg_student_mask_ratio": 0.465452536707744,
|
| 1497 |
+
"batch_ainp_frac": 0.0,
|
| 1498 |
+
"batch_inp_frac": 1.0,
|
| 1499 |
+
"batch_inp_oh_frac": 0.0,
|
| 1500 |
+
"batch_inp_par_frac": 0.0,
|
| 1501 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1502 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1503 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1504 |
+
"batch_rl_frac": 0.0,
|
| 1505 |
+
"batch_sft_frac": 0.0,
|
| 1506 |
+
"batch_soft_sft_frac": 0.0,
|
| 1507 |
+
"batch_tf_frac": 0.0,
|
| 1508 |
+
"ce_loss": 0.0,
|
| 1509 |
+
"epoch": 1.2389333333333332,
|
| 1510 |
+
"grad_norm": 0.251953125,
|
| 1511 |
+
"kd_loss": 0.44281044333342834,
|
| 1512 |
+
"learning_rate": 3e-06,
|
| 1513 |
+
"loss": 0.4225,
|
| 1514 |
+
"masked_tokens": 129.075,
|
| 1515 |
+
"mean_t": 0.49976949762785805,
|
| 1516 |
+
"step": 580,
|
| 1517 |
+
"student_masked_tokens": 129.075
|
| 1518 |
+
},
|
| 1519 |
+
{
|
| 1520 |
+
"avg_mask_ratio": 0.5225977989146486,
|
| 1521 |
+
"avg_response_length": 228.45,
|
| 1522 |
+
"avg_student_mask_ratio": 0.5225977989146486,
|
| 1523 |
+
"batch_ainp_frac": 0.0,
|
| 1524 |
+
"batch_inp_frac": 1.0,
|
| 1525 |
+
"batch_inp_oh_frac": 0.0,
|
| 1526 |
+
"batch_inp_par_frac": 0.0,
|
| 1527 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1528 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1529 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1530 |
+
"batch_rl_frac": 0.0,
|
| 1531 |
+
"batch_sft_frac": 0.0,
|
| 1532 |
+
"batch_soft_sft_frac": 0.0,
|
| 1533 |
+
"batch_tf_frac": 0.0,
|
| 1534 |
+
"ce_loss": 0.0,
|
| 1535 |
+
"epoch": 1.2602666666666666,
|
| 1536 |
+
"grad_norm": 0.251953125,
|
| 1537 |
+
"kd_loss": 0.5374341921963897,
|
| 1538 |
+
"learning_rate": 3e-06,
|
| 1539 |
+
"loss": 0.5561,
|
| 1540 |
+
"masked_tokens": 121.675,
|
| 1541 |
+
"mean_t": 0.5521843038732186,
|
| 1542 |
+
"step": 590,
|
| 1543 |
+
"student_masked_tokens": 121.675
|
| 1544 |
+
},
|
| 1545 |
+
{
|
| 1546 |
+
"avg_mask_ratio": 0.44998724836623294,
|
| 1547 |
+
"avg_response_length": 236.7,
|
| 1548 |
+
"avg_student_mask_ratio": 0.44998724836623294,
|
| 1549 |
+
"batch_ainp_frac": 0.0,
|
| 1550 |
+
"batch_inp_frac": 1.0,
|
| 1551 |
+
"batch_inp_oh_frac": 0.0,
|
| 1552 |
+
"batch_inp_par_frac": 0.0,
|
| 1553 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1554 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1555 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1556 |
+
"batch_rl_frac": 0.0,
|
| 1557 |
+
"batch_sft_frac": 0.0,
|
| 1558 |
+
"batch_soft_sft_frac": 0.0,
|
| 1559 |
+
"batch_tf_frac": 0.0,
|
| 1560 |
+
"ce_loss": 0.0,
|
| 1561 |
+
"epoch": 1.2816,
|
| 1562 |
+
"grad_norm": 0.2392578125,
|
| 1563 |
+
"kd_loss": 0.41848757546744225,
|
| 1564 |
+
"learning_rate": 3e-06,
|
| 1565 |
+
"loss": 0.4048,
|
| 1566 |
+
"masked_tokens": 110.1625,
|
| 1567 |
+
"mean_t": 0.4788527532829903,
|
| 1568 |
+
"step": 600,
|
| 1569 |
+
"student_masked_tokens": 110.1625
|
| 1570 |
+
},
|
| 1571 |
+
{
|
| 1572 |
+
"avg_mask_ratio": 0.5042130865273066,
|
| 1573 |
+
"avg_response_length": 230.3375,
|
| 1574 |
+
"avg_student_mask_ratio": 0.5042130865273066,
|
| 1575 |
+
"batch_ainp_frac": 0.0,
|
| 1576 |
+
"batch_inp_frac": 1.0,
|
| 1577 |
+
"batch_inp_oh_frac": 0.0,
|
| 1578 |
+
"batch_inp_par_frac": 0.0,
|
| 1579 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1580 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1581 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1582 |
+
"batch_rl_frac": 0.0,
|
| 1583 |
+
"batch_sft_frac": 0.0,
|
| 1584 |
+
"batch_soft_sft_frac": 0.0,
|
| 1585 |
+
"batch_tf_frac": 0.0,
|
| 1586 |
+
"ce_loss": 0.0,
|
| 1587 |
+
"epoch": 1.3029333333333333,
|
| 1588 |
+
"grad_norm": 0.255859375,
|
| 1589 |
+
"kd_loss": 0.5498438457787416,
|
| 1590 |
+
"learning_rate": 3e-06,
|
| 1591 |
+
"loss": 0.5621,
|
| 1592 |
+
"masked_tokens": 112.625,
|
| 1593 |
+
"mean_t": 0.5445419924799353,
|
| 1594 |
+
"step": 610,
|
| 1595 |
+
"student_masked_tokens": 112.625
|
| 1596 |
+
},
|
| 1597 |
+
{
|
| 1598 |
+
"avg_mask_ratio": 0.49637898594373836,
|
| 1599 |
+
"avg_response_length": 233.0625,
|
| 1600 |
+
"avg_student_mask_ratio": 0.49637898594373836,
|
| 1601 |
+
"batch_ainp_frac": 0.0,
|
| 1602 |
+
"batch_inp_frac": 1.0,
|
| 1603 |
+
"batch_inp_oh_frac": 0.0,
|
| 1604 |
+
"batch_inp_par_frac": 0.0,
|
| 1605 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1606 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1607 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1608 |
+
"batch_rl_frac": 0.0,
|
| 1609 |
+
"batch_sft_frac": 0.0,
|
| 1610 |
+
"batch_soft_sft_frac": 0.0,
|
| 1611 |
+
"batch_tf_frac": 0.0,
|
| 1612 |
+
"ce_loss": 0.0,
|
| 1613 |
+
"epoch": 1.3242666666666667,
|
| 1614 |
+
"grad_norm": 0.2060546875,
|
| 1615 |
+
"kd_loss": 0.5435912919167265,
|
| 1616 |
+
"learning_rate": 3e-06,
|
| 1617 |
+
"loss": 0.5191,
|
| 1618 |
+
"masked_tokens": 111.25,
|
| 1619 |
+
"mean_t": 0.52490478400141,
|
| 1620 |
+
"step": 620,
|
| 1621 |
+
"student_masked_tokens": 111.25
|
| 1622 |
+
},
|
| 1623 |
+
{
|
| 1624 |
+
"avg_mask_ratio": 0.5177568581304512,
|
| 1625 |
+
"avg_response_length": 257.2125,
|
| 1626 |
+
"avg_student_mask_ratio": 0.5177568581304512,
|
| 1627 |
+
"batch_ainp_frac": 0.0,
|
| 1628 |
+
"batch_inp_frac": 1.0,
|
| 1629 |
+
"batch_inp_oh_frac": 0.0,
|
| 1630 |
+
"batch_inp_par_frac": 0.0,
|
| 1631 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1632 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1633 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1634 |
+
"batch_rl_frac": 0.0,
|
| 1635 |
+
"batch_sft_frac": 0.0,
|
| 1636 |
+
"batch_soft_sft_frac": 0.0,
|
| 1637 |
+
"batch_tf_frac": 0.0,
|
| 1638 |
+
"ce_loss": 0.0,
|
| 1639 |
+
"epoch": 1.3456000000000001,
|
| 1640 |
+
"grad_norm": 0.423828125,
|
| 1641 |
+
"kd_loss": 0.534644598815612,
|
| 1642 |
+
"learning_rate": 3e-06,
|
| 1643 |
+
"loss": 0.5523,
|
| 1644 |
+
"masked_tokens": 129.6125,
|
| 1645 |
+
"mean_t": 0.5564947265549562,
|
| 1646 |
+
"step": 630,
|
| 1647 |
+
"student_masked_tokens": 129.6125
|
| 1648 |
+
},
|
| 1649 |
+
{
|
| 1650 |
+
"avg_mask_ratio": 0.48226998368045315,
|
| 1651 |
+
"avg_response_length": 237.7125,
|
| 1652 |
+
"avg_student_mask_ratio": 0.48226998368045315,
|
| 1653 |
+
"batch_ainp_frac": 0.0,
|
| 1654 |
+
"batch_inp_frac": 1.0,
|
| 1655 |
+
"batch_inp_oh_frac": 0.0,
|
| 1656 |
+
"batch_inp_par_frac": 0.0,
|
| 1657 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1658 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1659 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1660 |
+
"batch_rl_frac": 0.0,
|
| 1661 |
+
"batch_sft_frac": 0.0,
|
| 1662 |
+
"batch_soft_sft_frac": 0.0,
|
| 1663 |
+
"batch_tf_frac": 0.0,
|
| 1664 |
+
"ce_loss": 0.0,
|
| 1665 |
+
"epoch": 1.3669333333333333,
|
| 1666 |
+
"grad_norm": 0.138671875,
|
| 1667 |
+
"kd_loss": 0.363254164889031,
|
| 1668 |
+
"learning_rate": 3e-06,
|
| 1669 |
+
"loss": 0.3199,
|
| 1670 |
+
"masked_tokens": 120.275,
|
| 1671 |
+
"mean_t": 0.5093393943971023,
|
| 1672 |
+
"step": 640,
|
| 1673 |
+
"student_masked_tokens": 120.275
|
| 1674 |
+
},
|
| 1675 |
+
{
|
| 1676 |
+
"avg_mask_ratio": 0.5306948523037136,
|
| 1677 |
+
"avg_response_length": 238.0125,
|
| 1678 |
+
"avg_student_mask_ratio": 0.5306948523037136,
|
| 1679 |
+
"batch_ainp_frac": 0.0,
|
| 1680 |
+
"batch_inp_frac": 1.0,
|
| 1681 |
+
"batch_inp_oh_frac": 0.0,
|
| 1682 |
+
"batch_inp_par_frac": 0.0,
|
| 1683 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1684 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1685 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1686 |
+
"batch_rl_frac": 0.0,
|
| 1687 |
+
"batch_sft_frac": 0.0,
|
| 1688 |
+
"batch_soft_sft_frac": 0.0,
|
| 1689 |
+
"batch_tf_frac": 0.0,
|
| 1690 |
+
"ce_loss": 0.0,
|
| 1691 |
+
"epoch": 1.3882666666666665,
|
| 1692 |
+
"grad_norm": 0.158203125,
|
| 1693 |
+
"kd_loss": 0.5049424907911544,
|
| 1694 |
+
"learning_rate": 3e-06,
|
| 1695 |
+
"loss": 0.4932,
|
| 1696 |
+
"masked_tokens": 133.25,
|
| 1697 |
+
"mean_t": 0.5558586571365595,
|
| 1698 |
+
"step": 650,
|
| 1699 |
+
"student_masked_tokens": 133.25
|
| 1700 |
+
},
|
| 1701 |
+
{
|
| 1702 |
+
"avg_mask_ratio": 0.4821273953886703,
|
| 1703 |
+
"avg_response_length": 247.775,
|
| 1704 |
+
"avg_student_mask_ratio": 0.4821273953886703,
|
| 1705 |
+
"batch_ainp_frac": 0.0,
|
| 1706 |
+
"batch_inp_frac": 1.0,
|
| 1707 |
+
"batch_inp_oh_frac": 0.0,
|
| 1708 |
+
"batch_inp_par_frac": 0.0,
|
| 1709 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1710 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1711 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1712 |
+
"batch_rl_frac": 0.0,
|
| 1713 |
+
"batch_sft_frac": 0.0,
|
| 1714 |
+
"batch_soft_sft_frac": 0.0,
|
| 1715 |
+
"batch_tf_frac": 0.0,
|
| 1716 |
+
"ce_loss": 0.0,
|
| 1717 |
+
"epoch": 1.4096,
|
| 1718 |
+
"grad_norm": 0.23828125,
|
| 1719 |
+
"kd_loss": 0.41951912807553526,
|
| 1720 |
+
"learning_rate": 3e-06,
|
| 1721 |
+
"loss": 0.4242,
|
| 1722 |
+
"masked_tokens": 128.875,
|
| 1723 |
+
"mean_t": 0.51307404555846,
|
| 1724 |
+
"step": 660,
|
| 1725 |
+
"student_masked_tokens": 128.875
|
| 1726 |
+
},
|
| 1727 |
+
{
|
| 1728 |
+
"avg_mask_ratio": 0.46056515555246735,
|
| 1729 |
+
"avg_response_length": 240.4375,
|
| 1730 |
+
"avg_student_mask_ratio": 0.46056515555246735,
|
| 1731 |
+
"batch_ainp_frac": 0.0,
|
| 1732 |
+
"batch_inp_frac": 1.0,
|
| 1733 |
+
"batch_inp_oh_frac": 0.0,
|
| 1734 |
+
"batch_inp_par_frac": 0.0,
|
| 1735 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1736 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1737 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1738 |
+
"batch_rl_frac": 0.0,
|
| 1739 |
+
"batch_sft_frac": 0.0,
|
| 1740 |
+
"batch_soft_sft_frac": 0.0,
|
| 1741 |
+
"batch_tf_frac": 0.0,
|
| 1742 |
+
"ce_loss": 0.0,
|
| 1743 |
+
"epoch": 1.4309333333333334,
|
| 1744 |
+
"grad_norm": 0.28515625,
|
| 1745 |
+
"kd_loss": 0.3502249997708077,
|
| 1746 |
+
"learning_rate": 3e-06,
|
| 1747 |
+
"loss": 0.383,
|
| 1748 |
+
"masked_tokens": 119.85,
|
| 1749 |
+
"mean_t": 0.4907285622088239,
|
| 1750 |
+
"step": 670,
|
| 1751 |
+
"student_masked_tokens": 119.85
|
| 1752 |
+
},
|
| 1753 |
+
{
|
| 1754 |
+
"avg_mask_ratio": 0.4666106043441687,
|
| 1755 |
+
"avg_response_length": 226.7375,
|
| 1756 |
+
"avg_student_mask_ratio": 0.4666106043441687,
|
| 1757 |
+
"batch_ainp_frac": 0.0,
|
| 1758 |
+
"batch_inp_frac": 1.0,
|
| 1759 |
+
"batch_inp_oh_frac": 0.0,
|
| 1760 |
+
"batch_inp_par_frac": 0.0,
|
| 1761 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1762 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1763 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1764 |
+
"batch_rl_frac": 0.0,
|
| 1765 |
+
"batch_sft_frac": 0.0,
|
| 1766 |
+
"batch_soft_sft_frac": 0.0,
|
| 1767 |
+
"batch_tf_frac": 0.0,
|
| 1768 |
+
"ce_loss": 0.0,
|
| 1769 |
+
"epoch": 1.4522666666666666,
|
| 1770 |
+
"grad_norm": 0.3828125,
|
| 1771 |
+
"kd_loss": 0.4872104180909183,
|
| 1772 |
+
"learning_rate": 3e-06,
|
| 1773 |
+
"loss": 0.4906,
|
| 1774 |
+
"masked_tokens": 107.4625,
|
| 1775 |
+
"mean_t": 0.4913603452499956,
|
| 1776 |
+
"step": 680,
|
| 1777 |
+
"student_masked_tokens": 107.4625
|
| 1778 |
+
},
|
| 1779 |
+
{
|
| 1780 |
+
"avg_mask_ratio": 0.4790851596510038,
|
| 1781 |
+
"avg_response_length": 202.05,
|
| 1782 |
+
"avg_student_mask_ratio": 0.4790851596510038,
|
| 1783 |
+
"batch_ainp_frac": 0.0,
|
| 1784 |
+
"batch_inp_frac": 1.0,
|
| 1785 |
+
"batch_inp_oh_frac": 0.0,
|
| 1786 |
+
"batch_inp_par_frac": 0.0,
|
| 1787 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1788 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1789 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1790 |
+
"batch_rl_frac": 0.0,
|
| 1791 |
+
"batch_sft_frac": 0.0,
|
| 1792 |
+
"batch_soft_sft_frac": 0.0,
|
| 1793 |
+
"batch_tf_frac": 0.0,
|
| 1794 |
+
"ce_loss": 0.0,
|
| 1795 |
+
"epoch": 1.4736,
|
| 1796 |
+
"grad_norm": 0.2197265625,
|
| 1797 |
+
"kd_loss": 0.41381314450108847,
|
| 1798 |
+
"learning_rate": 3e-06,
|
| 1799 |
+
"loss": 0.453,
|
| 1800 |
+
"masked_tokens": 111.3125,
|
| 1801 |
+
"mean_t": 0.5133644798654131,
|
| 1802 |
+
"step": 690,
|
| 1803 |
+
"student_masked_tokens": 111.3125
|
| 1804 |
+
},
|
| 1805 |
+
{
|
| 1806 |
+
"avg_mask_ratio": 0.5250519359949977,
|
| 1807 |
+
"avg_response_length": 228.125,
|
| 1808 |
+
"avg_student_mask_ratio": 0.5250519359949977,
|
| 1809 |
+
"batch_ainp_frac": 0.0,
|
| 1810 |
+
"batch_inp_frac": 1.0,
|
| 1811 |
+
"batch_inp_oh_frac": 0.0,
|
| 1812 |
+
"batch_inp_par_frac": 0.0,
|
| 1813 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1814 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1815 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1816 |
+
"batch_rl_frac": 0.0,
|
| 1817 |
+
"batch_sft_frac": 0.0,
|
| 1818 |
+
"batch_soft_sft_frac": 0.0,
|
| 1819 |
+
"batch_tf_frac": 0.0,
|
| 1820 |
+
"ce_loss": 0.0,
|
| 1821 |
+
"epoch": 1.4949333333333334,
|
| 1822 |
+
"grad_norm": 0.2041015625,
|
| 1823 |
+
"kd_loss": 0.658003289926637,
|
| 1824 |
+
"learning_rate": 3e-06,
|
| 1825 |
+
"loss": 0.5988,
|
| 1826 |
+
"masked_tokens": 132.55,
|
| 1827 |
+
"mean_t": 0.5625698395539075,
|
| 1828 |
+
"step": 700,
|
| 1829 |
+
"student_masked_tokens": 132.55
|
| 1830 |
+
},
|
| 1831 |
+
{
|
| 1832 |
+
"avg_mask_ratio": 0.4790433386107907,
|
| 1833 |
+
"avg_response_length": 212.5,
|
| 1834 |
+
"avg_student_mask_ratio": 0.4790433386107907,
|
| 1835 |
+
"batch_ainp_frac": 0.0,
|
| 1836 |
+
"batch_inp_frac": 1.0,
|
| 1837 |
+
"batch_inp_oh_frac": 0.0,
|
| 1838 |
+
"batch_inp_par_frac": 0.0,
|
| 1839 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1840 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1841 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1842 |
+
"batch_rl_frac": 0.0,
|
| 1843 |
+
"batch_sft_frac": 0.0,
|
| 1844 |
+
"batch_soft_sft_frac": 0.0,
|
| 1845 |
+
"batch_tf_frac": 0.0,
|
| 1846 |
+
"ce_loss": 0.0,
|
| 1847 |
+
"epoch": 1.5162666666666667,
|
| 1848 |
+
"grad_norm": 0.173828125,
|
| 1849 |
+
"kd_loss": 0.42669102203761894,
|
| 1850 |
+
"learning_rate": 3e-06,
|
| 1851 |
+
"loss": 0.4103,
|
| 1852 |
+
"masked_tokens": 108.7375,
|
| 1853 |
+
"mean_t": 0.5135623761918395,
|
| 1854 |
+
"step": 710,
|
| 1855 |
+
"student_masked_tokens": 108.7375
|
| 1856 |
+
},
|
| 1857 |
+
{
|
| 1858 |
+
"avg_mask_ratio": 0.47950589570682495,
|
| 1859 |
+
"avg_response_length": 227.075,
|
| 1860 |
+
"avg_student_mask_ratio": 0.47950589570682495,
|
| 1861 |
+
"batch_ainp_frac": 0.0,
|
| 1862 |
+
"batch_inp_frac": 1.0,
|
| 1863 |
+
"batch_inp_oh_frac": 0.0,
|
| 1864 |
+
"batch_inp_par_frac": 0.0,
|
| 1865 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1866 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1867 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1868 |
+
"batch_rl_frac": 0.0,
|
| 1869 |
+
"batch_sft_frac": 0.0,
|
| 1870 |
+
"batch_soft_sft_frac": 0.0,
|
| 1871 |
+
"batch_tf_frac": 0.0,
|
| 1872 |
+
"ce_loss": 0.0,
|
| 1873 |
+
"epoch": 1.5375999999999999,
|
| 1874 |
+
"grad_norm": 0.365234375,
|
| 1875 |
+
"kd_loss": 0.562004916462638,
|
| 1876 |
+
"learning_rate": 3e-06,
|
| 1877 |
+
"loss": 0.5257,
|
| 1878 |
+
"masked_tokens": 110.8,
|
| 1879 |
+
"mean_t": 0.5117021896177902,
|
| 1880 |
+
"step": 720,
|
| 1881 |
+
"student_masked_tokens": 110.8
|
| 1882 |
+
},
|
| 1883 |
+
{
|
| 1884 |
+
"avg_mask_ratio": 0.4604924251558259,
|
| 1885 |
+
"avg_response_length": 232.925,
|
| 1886 |
+
"avg_student_mask_ratio": 0.4604924251558259,
|
| 1887 |
+
"batch_ainp_frac": 0.0,
|
| 1888 |
+
"batch_inp_frac": 1.0,
|
| 1889 |
+
"batch_inp_oh_frac": 0.0,
|
| 1890 |
+
"batch_inp_par_frac": 0.0,
|
| 1891 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1892 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1893 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1894 |
+
"batch_rl_frac": 0.0,
|
| 1895 |
+
"batch_sft_frac": 0.0,
|
| 1896 |
+
"batch_soft_sft_frac": 0.0,
|
| 1897 |
+
"batch_tf_frac": 0.0,
|
| 1898 |
+
"ce_loss": 0.0,
|
| 1899 |
+
"epoch": 1.5589333333333333,
|
| 1900 |
+
"grad_norm": 0.1630859375,
|
| 1901 |
+
"kd_loss": 0.42463611174693483,
|
| 1902 |
+
"learning_rate": 3e-06,
|
| 1903 |
+
"loss": 0.4182,
|
| 1904 |
+
"masked_tokens": 104.9625,
|
| 1905 |
+
"mean_t": 0.49050743713742123,
|
| 1906 |
+
"step": 730,
|
| 1907 |
+
"student_masked_tokens": 104.9625
|
| 1908 |
+
},
|
| 1909 |
+
{
|
| 1910 |
+
"avg_mask_ratio": 0.5185885130194947,
|
| 1911 |
+
"avg_response_length": 183.325,
|
| 1912 |
+
"avg_student_mask_ratio": 0.5185885130194947,
|
| 1913 |
+
"batch_ainp_frac": 0.0,
|
| 1914 |
+
"batch_inp_frac": 1.0,
|
| 1915 |
+
"batch_inp_oh_frac": 0.0,
|
| 1916 |
+
"batch_inp_par_frac": 0.0,
|
| 1917 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1918 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1919 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1920 |
+
"batch_rl_frac": 0.0,
|
| 1921 |
+
"batch_sft_frac": 0.0,
|
| 1922 |
+
"batch_soft_sft_frac": 0.0,
|
| 1923 |
+
"batch_tf_frac": 0.0,
|
| 1924 |
+
"ce_loss": 0.0,
|
| 1925 |
+
"epoch": 1.5802666666666667,
|
| 1926 |
+
"grad_norm": 0.1298828125,
|
| 1927 |
+
"kd_loss": 0.5289002722447321,
|
| 1928 |
+
"learning_rate": 3e-06,
|
| 1929 |
+
"loss": 0.5279,
|
| 1930 |
+
"masked_tokens": 97.125,
|
| 1931 |
+
"mean_t": 0.5505168779753149,
|
| 1932 |
+
"step": 740,
|
| 1933 |
+
"student_masked_tokens": 97.125
|
| 1934 |
+
},
|
| 1935 |
+
{
|
| 1936 |
+
"avg_mask_ratio": 0.4191439319110941,
|
| 1937 |
+
"avg_response_length": 223.65,
|
| 1938 |
+
"avg_student_mask_ratio": 0.4191439319110941,
|
| 1939 |
+
"batch_ainp_frac": 0.0,
|
| 1940 |
+
"batch_inp_frac": 1.0,
|
| 1941 |
+
"batch_inp_oh_frac": 0.0,
|
| 1942 |
+
"batch_inp_par_frac": 0.0,
|
| 1943 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1944 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1945 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1946 |
+
"batch_rl_frac": 0.0,
|
| 1947 |
+
"batch_sft_frac": 0.0,
|
| 1948 |
+
"batch_soft_sft_frac": 0.0,
|
| 1949 |
+
"batch_tf_frac": 0.0,
|
| 1950 |
+
"ce_loss": 0.0,
|
| 1951 |
+
"epoch": 1.6016,
|
| 1952 |
+
"grad_norm": 0.271484375,
|
| 1953 |
+
"kd_loss": 0.49676298911740346,
|
| 1954 |
+
"learning_rate": 3e-06,
|
| 1955 |
+
"loss": 0.4776,
|
| 1956 |
+
"masked_tokens": 95.3125,
|
| 1957 |
+
"mean_t": 0.4437690361432033,
|
| 1958 |
+
"step": 750,
|
| 1959 |
+
"student_masked_tokens": 95.3125
|
| 1960 |
+
},
|
| 1961 |
+
{
|
| 1962 |
+
"avg_mask_ratio": 0.46706983938929625,
|
| 1963 |
+
"avg_response_length": 216.0625,
|
| 1964 |
+
"avg_student_mask_ratio": 0.46706983938929625,
|
| 1965 |
+
"batch_ainp_frac": 0.0,
|
| 1966 |
+
"batch_inp_frac": 1.0,
|
| 1967 |
+
"batch_inp_oh_frac": 0.0,
|
| 1968 |
+
"batch_inp_par_frac": 0.0,
|
| 1969 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1970 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1971 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1972 |
+
"batch_rl_frac": 0.0,
|
| 1973 |
+
"batch_sft_frac": 0.0,
|
| 1974 |
+
"batch_soft_sft_frac": 0.0,
|
| 1975 |
+
"batch_tf_frac": 0.0,
|
| 1976 |
+
"ce_loss": 0.0,
|
| 1977 |
+
"epoch": 1.6229333333333333,
|
| 1978 |
+
"grad_norm": 1.234375,
|
| 1979 |
+
"kd_loss": 0.5423888156364214,
|
| 1980 |
+
"learning_rate": 3e-06,
|
| 1981 |
+
"loss": 0.5307,
|
| 1982 |
+
"masked_tokens": 107.7,
|
| 1983 |
+
"mean_t": 0.49132869170280175,
|
| 1984 |
+
"step": 760,
|
| 1985 |
+
"student_masked_tokens": 107.7
|
| 1986 |
+
},
|
| 1987 |
+
{
|
| 1988 |
+
"avg_mask_ratio": 0.44156218122225255,
|
| 1989 |
+
"avg_response_length": 259.675,
|
| 1990 |
+
"avg_student_mask_ratio": 0.44156218122225255,
|
| 1991 |
+
"batch_ainp_frac": 0.0,
|
| 1992 |
+
"batch_inp_frac": 1.0,
|
| 1993 |
+
"batch_inp_oh_frac": 0.0,
|
| 1994 |
+
"batch_inp_par_frac": 0.0,
|
| 1995 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1996 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1997 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1998 |
+
"batch_rl_frac": 0.0,
|
| 1999 |
+
"batch_sft_frac": 0.0,
|
| 2000 |
+
"batch_soft_sft_frac": 0.0,
|
| 2001 |
+
"batch_tf_frac": 0.0,
|
| 2002 |
+
"ce_loss": 0.0,
|
| 2003 |
+
"epoch": 1.6442666666666668,
|
| 2004 |
+
"grad_norm": 0.1455078125,
|
| 2005 |
+
"kd_loss": 0.4092745722692257,
|
| 2006 |
+
"learning_rate": 3e-06,
|
| 2007 |
+
"loss": 0.3866,
|
| 2008 |
+
"masked_tokens": 117.5,
|
| 2009 |
+
"mean_t": 0.4667695587326307,
|
| 2010 |
+
"step": 770,
|
| 2011 |
+
"student_masked_tokens": 117.5
|
| 2012 |
+
},
|
| 2013 |
+
{
|
| 2014 |
+
"avg_mask_ratio": 0.42836043585848527,
|
| 2015 |
+
"avg_response_length": 258.5125,
|
| 2016 |
+
"avg_student_mask_ratio": 0.42836043585848527,
|
| 2017 |
+
"batch_ainp_frac": 0.0,
|
| 2018 |
+
"batch_inp_frac": 1.0,
|
| 2019 |
+
"batch_inp_oh_frac": 0.0,
|
| 2020 |
+
"batch_inp_par_frac": 0.0,
|
| 2021 |
+
"batch_inp_par_par_frac": 0.0,
|
| 2022 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 2023 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 2024 |
+
"batch_rl_frac": 0.0,
|
| 2025 |
+
"batch_sft_frac": 0.0,
|
| 2026 |
+
"batch_soft_sft_frac": 0.0,
|
| 2027 |
+
"batch_tf_frac": 0.0,
|
| 2028 |
+
"ce_loss": 0.0,
|
| 2029 |
+
"epoch": 1.6656,
|
| 2030 |
+
"grad_norm": 0.216796875,
|
| 2031 |
+
"kd_loss": 0.3423420360170297,
|
| 2032 |
+
"learning_rate": 3e-06,
|
| 2033 |
+
"loss": 0.3535,
|
| 2034 |
+
"masked_tokens": 119.0125,
|
| 2035 |
+
"mean_t": 0.44942845597106496,
|
| 2036 |
+
"step": 780,
|
| 2037 |
+
"student_masked_tokens": 119.0125
|
| 2038 |
+
},
|
| 2039 |
+
{
|
| 2040 |
+
"avg_mask_ratio": 0.46589430308085866,
|
| 2041 |
+
"avg_response_length": 222.3125,
|
| 2042 |
+
"avg_student_mask_ratio": 0.46589430308085866,
|
| 2043 |
+
"batch_ainp_frac": 0.0,
|
| 2044 |
+
"batch_inp_frac": 1.0,
|
| 2045 |
+
"batch_inp_oh_frac": 0.0,
|
| 2046 |
+
"batch_inp_par_frac": 0.0,
|
| 2047 |
+
"batch_inp_par_par_frac": 0.0,
|
| 2048 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 2049 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 2050 |
+
"batch_rl_frac": 0.0,
|
| 2051 |
+
"batch_sft_frac": 0.0,
|
| 2052 |
+
"batch_soft_sft_frac": 0.0,
|
| 2053 |
+
"batch_tf_frac": 0.0,
|
| 2054 |
+
"ce_loss": 0.0,
|
| 2055 |
+
"epoch": 1.6869333333333332,
|
| 2056 |
+
"grad_norm": 0.125,
|
| 2057 |
+
"kd_loss": 0.33727191750285523,
|
| 2058 |
+
"learning_rate": 3e-06,
|
| 2059 |
+
"loss": 0.3767,
|
| 2060 |
+
"masked_tokens": 103.25,
|
| 2061 |
+
"mean_t": 0.4993515375303105,
|
| 2062 |
+
"step": 790,
|
| 2063 |
+
"student_masked_tokens": 103.25
|
| 2064 |
+
},
|
| 2065 |
+
{
|
| 2066 |
+
"avg_mask_ratio": 0.46366424662992356,
|
| 2067 |
+
"avg_response_length": 219.6875,
|
| 2068 |
+
"avg_student_mask_ratio": 0.46366424662992356,
|
| 2069 |
+
"batch_ainp_frac": 0.0,
|
| 2070 |
+
"batch_inp_frac": 1.0,
|
| 2071 |
+
"batch_inp_oh_frac": 0.0,
|
| 2072 |
+
"batch_inp_par_frac": 0.0,
|
| 2073 |
+
"batch_inp_par_par_frac": 0.0,
|
| 2074 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 2075 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 2076 |
+
"batch_rl_frac": 0.0,
|
| 2077 |
+
"batch_sft_frac": 0.0,
|
| 2078 |
+
"batch_soft_sft_frac": 0.0,
|
| 2079 |
+
"batch_tf_frac": 0.0,
|
| 2080 |
+
"ce_loss": 0.0,
|
| 2081 |
+
"epoch": 1.7082666666666668,
|
| 2082 |
+
"grad_norm": 0.189453125,
|
| 2083 |
+
"kd_loss": 0.34817896180302343,
|
| 2084 |
+
"learning_rate": 3e-06,
|
| 2085 |
+
"loss": 0.3863,
|
| 2086 |
+
"masked_tokens": 104.5,
|
| 2087 |
+
"mean_t": 0.500370389316231,
|
| 2088 |
+
"step": 800,
|
| 2089 |
+
"student_masked_tokens": 104.5
|
| 2090 |
+
},
|
| 2091 |
+
{
|
| 2092 |
+
"avg_mask_ratio": 0.503375941584818,
|
| 2093 |
+
"avg_response_length": 237.85,
|
| 2094 |
+
"avg_student_mask_ratio": 0.503375941584818,
|
| 2095 |
+
"batch_ainp_frac": 0.0,
|
| 2096 |
+
"batch_inp_frac": 1.0,
|
| 2097 |
+
"batch_inp_oh_frac": 0.0,
|
| 2098 |
+
"batch_inp_par_frac": 0.0,
|
| 2099 |
+
"batch_inp_par_par_frac": 0.0,
|
| 2100 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 2101 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 2102 |
+
"batch_rl_frac": 0.0,
|
| 2103 |
+
"batch_sft_frac": 0.0,
|
| 2104 |
+
"batch_soft_sft_frac": 0.0,
|
| 2105 |
+
"batch_tf_frac": 0.0,
|
| 2106 |
+
"ce_loss": 0.0,
|
| 2107 |
+
"epoch": 1.7296,
|
| 2108 |
+
"grad_norm": 0.234375,
|
| 2109 |
+
"kd_loss": 0.4517137570231853,
|
| 2110 |
+
"learning_rate": 3e-06,
|
| 2111 |
+
"loss": 0.464,
|
| 2112 |
+
"masked_tokens": 118.2125,
|
| 2113 |
+
"mean_t": 0.5317009104182944,
|
| 2114 |
+
"step": 810,
|
| 2115 |
+
"student_masked_tokens": 118.2125
|
| 2116 |
+
},
|
| 2117 |
+
{
|
| 2118 |
+
"avg_mask_ratio": 0.5110091455746442,
|
| 2119 |
+
"avg_response_length": 209.0875,
|
| 2120 |
+
"avg_student_mask_ratio": 0.5110091455746442,
|
| 2121 |
+
"batch_ainp_frac": 0.0,
|
| 2122 |
+
"batch_inp_frac": 1.0,
|
| 2123 |
+
"batch_inp_oh_frac": 0.0,
|
| 2124 |
+
"batch_inp_par_frac": 0.0,
|
| 2125 |
+
"batch_inp_par_par_frac": 0.0,
|
| 2126 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 2127 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 2128 |
+
"batch_rl_frac": 0.0,
|
| 2129 |
+
"batch_sft_frac": 0.0,
|
| 2130 |
+
"batch_soft_sft_frac": 0.0,
|
| 2131 |
+
"batch_tf_frac": 0.0,
|
| 2132 |
+
"ce_loss": 0.0,
|
| 2133 |
+
"epoch": 1.7509333333333332,
|
| 2134 |
+
"grad_norm": 0.33203125,
|
| 2135 |
+
"kd_loss": 0.590595377480895,
|
| 2136 |
+
"learning_rate": 3e-06,
|
| 2137 |
+
"loss": 0.5476,
|
| 2138 |
+
"masked_tokens": 120.95,
|
| 2139 |
+
"mean_t": 0.5392061032878701,
|
| 2140 |
+
"step": 820,
|
| 2141 |
+
"student_masked_tokens": 120.95
|
| 2142 |
+
},
|
| 2143 |
+
{
|
| 2144 |
+
"avg_mask_ratio": 0.49899387182667854,
|
| 2145 |
+
"avg_response_length": 263.975,
|
| 2146 |
+
"avg_student_mask_ratio": 0.49899387182667854,
|
| 2147 |
+
"batch_ainp_frac": 0.0,
|
| 2148 |
+
"batch_inp_frac": 1.0,
|
| 2149 |
+
"batch_inp_oh_frac": 0.0,
|
| 2150 |
+
"batch_inp_par_frac": 0.0,
|
| 2151 |
+
"batch_inp_par_par_frac": 0.0,
|
| 2152 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 2153 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 2154 |
+
"batch_rl_frac": 0.0,
|
| 2155 |
+
"batch_sft_frac": 0.0,
|
| 2156 |
+
"batch_soft_sft_frac": 0.0,
|
| 2157 |
+
"batch_tf_frac": 0.0,
|
| 2158 |
+
"ce_loss": 0.0,
|
| 2159 |
+
"epoch": 1.7722666666666667,
|
| 2160 |
+
"grad_norm": 0.1455078125,
|
| 2161 |
+
"kd_loss": 0.5558899336714603,
|
| 2162 |
+
"learning_rate": 3e-06,
|
| 2163 |
+
"loss": 0.5794,
|
| 2164 |
+
"masked_tokens": 137.075,
|
| 2165 |
+
"mean_t": 0.5238314627087675,
|
| 2166 |
+
"step": 830,
|
| 2167 |
+
"student_masked_tokens": 137.075
|
| 2168 |
+
},
|
| 2169 |
+
{
|
| 2170 |
+
"avg_mask_ratio": 0.4997270987310912,
|
| 2171 |
+
"avg_response_length": 221.9,
|
| 2172 |
+
"avg_student_mask_ratio": 0.4997270987310912,
|
| 2173 |
+
"batch_ainp_frac": 0.0,
|
| 2174 |
+
"batch_inp_frac": 1.0,
|
| 2175 |
+
"batch_inp_oh_frac": 0.0,
|
| 2176 |
+
"batch_inp_par_frac": 0.0,
|
| 2177 |
+
"batch_inp_par_par_frac": 0.0,
|
| 2178 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 2179 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 2180 |
+
"batch_rl_frac": 0.0,
|
| 2181 |
+
"batch_sft_frac": 0.0,
|
| 2182 |
+
"batch_soft_sft_frac": 0.0,
|
| 2183 |
+
"batch_tf_frac": 0.0,
|
| 2184 |
+
"ce_loss": 0.0,
|
| 2185 |
+
"epoch": 1.7936,
|
| 2186 |
+
"grad_norm": 0.126953125,
|
| 2187 |
+
"kd_loss": 0.4676562640790394,
|
| 2188 |
+
"learning_rate": 3e-06,
|
| 2189 |
+
"loss": 0.4654,
|
| 2190 |
+
"masked_tokens": 114.525,
|
| 2191 |
+
"mean_t": 0.5301066277665086,
|
| 2192 |
+
"step": 840,
|
| 2193 |
+
"student_masked_tokens": 114.525
|
| 2194 |
+
},
|
| 2195 |
+
{
|
| 2196 |
+
"avg_mask_ratio": 0.4988076956477016,
|
| 2197 |
+
"avg_response_length": 225.5,
|
| 2198 |
+
"avg_student_mask_ratio": 0.4988076956477016,
|
| 2199 |
+
"batch_ainp_frac": 0.0,
|
| 2200 |
+
"batch_inp_frac": 1.0,
|
| 2201 |
+
"batch_inp_oh_frac": 0.0,
|
| 2202 |
+
"batch_inp_par_frac": 0.0,
|
| 2203 |
+
"batch_inp_par_par_frac": 0.0,
|
| 2204 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 2205 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 2206 |
+
"batch_rl_frac": 0.0,
|
| 2207 |
+
"batch_sft_frac": 0.0,
|
| 2208 |
+
"batch_soft_sft_frac": 0.0,
|
| 2209 |
+
"batch_tf_frac": 0.0,
|
| 2210 |
+
"ce_loss": 0.0,
|
| 2211 |
+
"epoch": 1.8149333333333333,
|
| 2212 |
+
"grad_norm": 0.11474609375,
|
| 2213 |
+
"kd_loss": 0.3244839489429069,
|
| 2214 |
+
"learning_rate": 3e-06,
|
| 2215 |
+
"loss": 0.3941,
|
| 2216 |
+
"masked_tokens": 116.9125,
|
| 2217 |
+
"mean_t": 0.5343429344706238,
|
| 2218 |
+
"step": 850,
|
| 2219 |
+
"student_masked_tokens": 116.9125
|
| 2220 |
+
},
|
| 2221 |
+
{
|
| 2222 |
+
"avg_mask_ratio": 0.4635998342186213,
|
| 2223 |
+
"avg_response_length": 229.125,
|
| 2224 |
+
"avg_student_mask_ratio": 0.4635998342186213,
|
| 2225 |
+
"batch_ainp_frac": 0.0,
|
| 2226 |
+
"batch_inp_frac": 1.0,
|
| 2227 |
+
"batch_inp_oh_frac": 0.0,
|
| 2228 |
+
"batch_inp_par_frac": 0.0,
|
| 2229 |
+
"batch_inp_par_par_frac": 0.0,
|
| 2230 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 2231 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 2232 |
+
"batch_rl_frac": 0.0,
|
| 2233 |
+
"batch_sft_frac": 0.0,
|
| 2234 |
+
"batch_soft_sft_frac": 0.0,
|
| 2235 |
+
"batch_tf_frac": 0.0,
|
| 2236 |
+
"ce_loss": 0.0,
|
| 2237 |
+
"epoch": 1.8362666666666667,
|
| 2238 |
+
"grad_norm": 0.181640625,
|
| 2239 |
+
"kd_loss": 0.4306980623789968,
|
| 2240 |
+
"learning_rate": 3e-06,
|
| 2241 |
+
"loss": 0.4216,
|
| 2242 |
+
"masked_tokens": 109.675,
|
| 2243 |
+
"mean_t": 0.4791536889737472,
|
| 2244 |
+
"step": 860,
|
| 2245 |
+
"student_masked_tokens": 109.675
|
| 2246 |
+
},
|
| 2247 |
+
{
|
| 2248 |
+
"avg_mask_ratio": 0.49111039767740294,
|
| 2249 |
+
"avg_response_length": 229.1,
|
| 2250 |
+
"avg_student_mask_ratio": 0.49111039767740294,
|
| 2251 |
+
"batch_ainp_frac": 0.0,
|
| 2252 |
+
"batch_inp_frac": 1.0,
|
| 2253 |
+
"batch_inp_oh_frac": 0.0,
|
| 2254 |
+
"batch_inp_par_frac": 0.0,
|
| 2255 |
+
"batch_inp_par_par_frac": 0.0,
|
| 2256 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 2257 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 2258 |
+
"batch_rl_frac": 0.0,
|
| 2259 |
+
"batch_sft_frac": 0.0,
|
| 2260 |
+
"batch_soft_sft_frac": 0.0,
|
| 2261 |
+
"batch_tf_frac": 0.0,
|
| 2262 |
+
"ce_loss": 0.0,
|
| 2263 |
+
"epoch": 1.8576000000000001,
|
| 2264 |
+
"grad_norm": 0.2412109375,
|
| 2265 |
+
"kd_loss": 0.4694033406512972,
|
| 2266 |
+
"learning_rate": 3e-06,
|
| 2267 |
+
"loss": 0.5022,
|
| 2268 |
+
"masked_tokens": 115.6625,
|
| 2269 |
+
"mean_t": 0.5203817339061061,
|
| 2270 |
+
"step": 870,
|
| 2271 |
+
"student_masked_tokens": 115.6625
|
| 2272 |
+
},
|
| 2273 |
+
{
|
| 2274 |
+
"avg_mask_ratio": 0.4427660425659269,
|
| 2275 |
+
"avg_response_length": 198.5625,
|
| 2276 |
+
"avg_student_mask_ratio": 0.4427660425659269,
|
| 2277 |
+
"batch_ainp_frac": 0.0,
|
| 2278 |
+
"batch_inp_frac": 1.0,
|
| 2279 |
+
"batch_inp_oh_frac": 0.0,
|
| 2280 |
+
"batch_inp_par_frac": 0.0,
|
| 2281 |
+
"batch_inp_par_par_frac": 0.0,
|
| 2282 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 2283 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 2284 |
+
"batch_rl_frac": 0.0,
|
| 2285 |
+
"batch_sft_frac": 0.0,
|
| 2286 |
+
"batch_soft_sft_frac": 0.0,
|
| 2287 |
+
"batch_tf_frac": 0.0,
|
| 2288 |
+
"ce_loss": 0.0,
|
| 2289 |
+
"epoch": 1.8789333333333333,
|
| 2290 |
+
"grad_norm": 0.1728515625,
|
| 2291 |
+
"kd_loss": 0.46008693509684234,
|
| 2292 |
+
"learning_rate": 3e-06,
|
| 2293 |
+
"loss": 0.4582,
|
| 2294 |
+
"masked_tokens": 91.175,
|
| 2295 |
+
"mean_t": 0.4875184997683391,
|
| 2296 |
+
"step": 880,
|
| 2297 |
+
"student_masked_tokens": 91.175
|
| 2298 |
+
},
|
| 2299 |
+
{
|
| 2300 |
+
"avg_mask_ratio": 0.4464349385118112,
|
| 2301 |
+
"avg_response_length": 225.8375,
|
| 2302 |
+
"avg_student_mask_ratio": 0.4464349385118112,
|
| 2303 |
+
"batch_ainp_frac": 0.0,
|
| 2304 |
+
"batch_inp_frac": 1.0,
|
| 2305 |
+
"batch_inp_oh_frac": 0.0,
|
| 2306 |
+
"batch_inp_par_frac": 0.0,
|
| 2307 |
+
"batch_inp_par_par_frac": 0.0,
|
| 2308 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 2309 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 2310 |
+
"batch_rl_frac": 0.0,
|
| 2311 |
+
"batch_sft_frac": 0.0,
|
| 2312 |
+
"batch_soft_sft_frac": 0.0,
|
| 2313 |
+
"batch_tf_frac": 0.0,
|
| 2314 |
+
"ce_loss": 0.0,
|
| 2315 |
+
"epoch": 1.9002666666666665,
|
| 2316 |
+
"grad_norm": 0.0888671875,
|
| 2317 |
+
"kd_loss": 0.3971562616849951,
|
| 2318 |
+
"learning_rate": 3e-06,
|
| 2319 |
+
"loss": 0.3817,
|
| 2320 |
+
"masked_tokens": 101.75,
|
| 2321 |
+
"mean_t": 0.4766692223958671,
|
| 2322 |
+
"step": 890,
|
| 2323 |
+
"student_masked_tokens": 101.75
|
| 2324 |
+
},
|
| 2325 |
+
{
|
| 2326 |
+
"avg_mask_ratio": 0.44976164362160487,
|
| 2327 |
+
"avg_response_length": 227.7875,
|
| 2328 |
+
"avg_student_mask_ratio": 0.44976164362160487,
|
| 2329 |
+
"batch_ainp_frac": 0.0,
|
| 2330 |
+
"batch_inp_frac": 1.0,
|
| 2331 |
+
"batch_inp_oh_frac": 0.0,
|
| 2332 |
+
"batch_inp_par_frac": 0.0,
|
| 2333 |
+
"batch_inp_par_par_frac": 0.0,
|
| 2334 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 2335 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 2336 |
+
"batch_rl_frac": 0.0,
|
| 2337 |
+
"batch_sft_frac": 0.0,
|
| 2338 |
+
"batch_soft_sft_frac": 0.0,
|
| 2339 |
+
"batch_tf_frac": 0.0,
|
| 2340 |
+
"ce_loss": 0.0,
|
| 2341 |
+
"epoch": 1.9216,
|
| 2342 |
+
"grad_norm": 0.1708984375,
|
| 2343 |
+
"kd_loss": 0.4591321104033284,
|
| 2344 |
+
"learning_rate": 3e-06,
|
| 2345 |
+
"loss": 0.4498,
|
| 2346 |
+
"masked_tokens": 103.475,
|
| 2347 |
+
"mean_t": 0.487134758150205,
|
| 2348 |
+
"step": 900,
|
| 2349 |
+
"student_masked_tokens": 103.475
|
| 2350 |
+
},
|
| 2351 |
+
{
|
| 2352 |
+
"avg_mask_ratio": 0.475579984736396,
|
| 2353 |
+
"avg_response_length": 245.1625,
|
| 2354 |
+
"avg_student_mask_ratio": 0.475579984736396,
|
| 2355 |
+
"batch_ainp_frac": 0.0,
|
| 2356 |
+
"batch_inp_frac": 1.0,
|
| 2357 |
+
"batch_inp_oh_frac": 0.0,
|
| 2358 |
+
"batch_inp_par_frac": 0.0,
|
| 2359 |
+
"batch_inp_par_par_frac": 0.0,
|
| 2360 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 2361 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 2362 |
+
"batch_rl_frac": 0.0,
|
| 2363 |
+
"batch_sft_frac": 0.0,
|
| 2364 |
+
"batch_soft_sft_frac": 0.0,
|
| 2365 |
+
"batch_tf_frac": 0.0,
|
| 2366 |
+
"ce_loss": 0.0,
|
| 2367 |
+
"epoch": 1.9429333333333334,
|
| 2368 |
+
"grad_norm": 0.1328125,
|
| 2369 |
+
"kd_loss": 0.4644632683706561,
|
| 2370 |
+
"learning_rate": 3e-06,
|
| 2371 |
+
"loss": 0.4135,
|
| 2372 |
+
"masked_tokens": 124.6375,
|
| 2373 |
+
"mean_t": 0.5027793228859082,
|
| 2374 |
+
"step": 910,
|
| 2375 |
+
"student_masked_tokens": 124.6375
|
| 2376 |
+
},
|
| 2377 |
+
{
|
| 2378 |
+
"avg_mask_ratio": 0.4688875659601763,
|
| 2379 |
+
"avg_response_length": 226.2875,
|
| 2380 |
+
"avg_student_mask_ratio": 0.4688875659601763,
|
| 2381 |
+
"batch_ainp_frac": 0.0,
|
| 2382 |
+
"batch_inp_frac": 1.0,
|
| 2383 |
+
"batch_inp_oh_frac": 0.0,
|
| 2384 |
+
"batch_inp_par_frac": 0.0,
|
| 2385 |
+
"batch_inp_par_par_frac": 0.0,
|
| 2386 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 2387 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 2388 |
+
"batch_rl_frac": 0.0,
|
| 2389 |
+
"batch_sft_frac": 0.0,
|
| 2390 |
+
"batch_soft_sft_frac": 0.0,
|
| 2391 |
+
"batch_tf_frac": 0.0,
|
| 2392 |
+
"ce_loss": 0.0,
|
| 2393 |
+
"epoch": 1.9642666666666666,
|
| 2394 |
+
"grad_norm": 0.2099609375,
|
| 2395 |
+
"kd_loss": 0.4406846639433752,
|
| 2396 |
+
"learning_rate": 3e-06,
|
| 2397 |
+
"loss": 0.4429,
|
| 2398 |
+
"masked_tokens": 110.0125,
|
| 2399 |
+
"mean_t": 0.49417946098838,
|
| 2400 |
+
"step": 920,
|
| 2401 |
+
"student_masked_tokens": 110.0125
|
| 2402 |
+
},
|
| 2403 |
+
{
|
| 2404 |
+
"avg_mask_ratio": 0.47999348094454036,
|
| 2405 |
+
"avg_response_length": 237.05,
|
| 2406 |
+
"avg_student_mask_ratio": 0.47999348094454036,
|
| 2407 |
+
"batch_ainp_frac": 0.0,
|
| 2408 |
+
"batch_inp_frac": 1.0,
|
| 2409 |
+
"batch_inp_oh_frac": 0.0,
|
| 2410 |
+
"batch_inp_par_frac": 0.0,
|
| 2411 |
+
"batch_inp_par_par_frac": 0.0,
|
| 2412 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 2413 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 2414 |
+
"batch_rl_frac": 0.0,
|
| 2415 |
+
"batch_sft_frac": 0.0,
|
| 2416 |
+
"batch_soft_sft_frac": 0.0,
|
| 2417 |
+
"batch_tf_frac": 0.0,
|
| 2418 |
+
"ce_loss": 0.0,
|
| 2419 |
+
"epoch": 1.9856,
|
| 2420 |
+
"grad_norm": 0.2275390625,
|
| 2421 |
+
"kd_loss": 0.4538009871917458,
|
| 2422 |
+
"learning_rate": 3e-06,
|
| 2423 |
+
"loss": 0.4455,
|
| 2424 |
+
"masked_tokens": 121.6,
|
| 2425 |
+
"mean_t": 0.5045580042526125,
|
| 2426 |
+
"step": 930,
|
| 2427 |
+
"student_masked_tokens": 121.6
|
| 2428 |
+
},
|
| 2429 |
+
{
|
| 2430 |
+
"avg_mask_ratio": 0.49413903727240505,
|
| 2431 |
+
"avg_response_length": 224.79761904761904,
|
| 2432 |
+
"avg_student_mask_ratio": 0.49413903727240505,
|
| 2433 |
+
"batch_ainp_frac": 0.0,
|
| 2434 |
+
"batch_inp_frac": 1.0,
|
| 2435 |
+
"batch_inp_oh_frac": 0.0,
|
| 2436 |
+
"batch_inp_par_frac": 0.0,
|
| 2437 |
+
"batch_inp_par_par_frac": 0.0,
|
| 2438 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 2439 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 2440 |
+
"batch_rl_frac": 0.0,
|
| 2441 |
+
"batch_sft_frac": 0.0,
|
| 2442 |
+
"batch_soft_sft_frac": 0.0,
|
| 2443 |
+
"batch_tf_frac": 0.0,
|
| 2444 |
+
"ce_loss": 0.0,
|
| 2445 |
+
"epoch": 2.0085333333333333,
|
| 2446 |
+
"grad_norm": 0.294921875,
|
| 2447 |
+
"kd_loss": 0.4914910109509676,
|
| 2448 |
+
"learning_rate": 3e-06,
|
| 2449 |
+
"loss": 0.4787,
|
| 2450 |
+
"masked_tokens": 120.5,
|
| 2451 |
+
"mean_t": 0.5321138524893849,
|
| 2452 |
+
"step": 940,
|
| 2453 |
+
"student_masked_tokens": 120.5
|
| 2454 |
+
},
|
| 2455 |
+
{
|
| 2456 |
+
"avg_mask_ratio": 0.4368605303927325,
|
| 2457 |
+
"avg_response_length": 240.9125,
|
| 2458 |
+
"avg_student_mask_ratio": 0.4368605303927325,
|
| 2459 |
+
"batch_ainp_frac": 0.0,
|
| 2460 |
+
"batch_inp_frac": 1.0,
|
| 2461 |
+
"batch_inp_oh_frac": 0.0,
|
| 2462 |
+
"batch_inp_par_frac": 0.0,
|
| 2463 |
+
"batch_inp_par_par_frac": 0.0,
|
| 2464 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 2465 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 2466 |
+
"batch_rl_frac": 0.0,
|
| 2467 |
+
"batch_sft_frac": 0.0,
|
| 2468 |
+
"batch_soft_sft_frac": 0.0,
|
| 2469 |
+
"batch_tf_frac": 0.0,
|
| 2470 |
+
"ce_loss": 0.0,
|
| 2471 |
+
"epoch": 2.0298666666666665,
|
| 2472 |
+
"grad_norm": 0.162109375,
|
| 2473 |
+
"kd_loss": 0.42506327110539815,
|
| 2474 |
+
"learning_rate": 3e-06,
|
| 2475 |
+
"loss": 0.4051,
|
| 2476 |
+
"masked_tokens": 111.4125,
|
| 2477 |
+
"mean_t": 0.4632946296595037,
|
| 2478 |
+
"step": 950,
|
| 2479 |
+
"student_masked_tokens": 111.4125
|
| 2480 |
+
},
|
| 2481 |
+
{
|
| 2482 |
+
"avg_mask_ratio": 0.4988762516761199,
|
| 2483 |
+
"avg_response_length": 275.3,
|
| 2484 |
+
"avg_student_mask_ratio": 0.4988762516761199,
|
| 2485 |
+
"batch_ainp_frac": 0.0,
|
| 2486 |
+
"batch_inp_frac": 1.0,
|
| 2487 |
+
"batch_inp_oh_frac": 0.0,
|
| 2488 |
+
"batch_inp_par_frac": 0.0,
|
| 2489 |
+
"batch_inp_par_par_frac": 0.0,
|
| 2490 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 2491 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 2492 |
+
"batch_rl_frac": 0.0,
|
| 2493 |
+
"batch_sft_frac": 0.0,
|
| 2494 |
+
"batch_soft_sft_frac": 0.0,
|
| 2495 |
+
"batch_tf_frac": 0.0,
|
| 2496 |
+
"ce_loss": 0.0,
|
| 2497 |
+
"epoch": 2.0512,
|
| 2498 |
+
"grad_norm": 0.1103515625,
|
| 2499 |
+
"kd_loss": 0.5790403748234894,
|
| 2500 |
+
"learning_rate": 3e-06,
|
| 2501 |
+
"loss": 0.5466,
|
| 2502 |
+
"masked_tokens": 143.825,
|
| 2503 |
+
"mean_t": 0.5198000721400604,
|
| 2504 |
+
"step": 960,
|
| 2505 |
+
"student_masked_tokens": 143.825
|
| 2506 |
+
},
|
| 2507 |
+
{
|
| 2508 |
+
"avg_mask_ratio": 0.437801384011982,
|
| 2509 |
+
"avg_response_length": 236.2375,
|
| 2510 |
+
"avg_student_mask_ratio": 0.437801384011982,
|
| 2511 |
+
"batch_ainp_frac": 0.0,
|
| 2512 |
+
"batch_inp_frac": 1.0,
|
| 2513 |
+
"batch_inp_oh_frac": 0.0,
|
| 2514 |
+
"batch_inp_par_frac": 0.0,
|
| 2515 |
+
"batch_inp_par_par_frac": 0.0,
|
| 2516 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 2517 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 2518 |
+
"batch_rl_frac": 0.0,
|
| 2519 |
+
"batch_sft_frac": 0.0,
|
| 2520 |
+
"batch_soft_sft_frac": 0.0,
|
| 2521 |
+
"batch_tf_frac": 0.0,
|
| 2522 |
+
"ce_loss": 0.0,
|
| 2523 |
+
"epoch": 2.0725333333333333,
|
| 2524 |
+
"grad_norm": 0.248046875,
|
| 2525 |
+
"kd_loss": 0.35248266551700097,
|
| 2526 |
+
"learning_rate": 3e-06,
|
| 2527 |
+
"loss": 0.3947,
|
| 2528 |
+
"masked_tokens": 107.025,
|
| 2529 |
+
"mean_t": 0.4703940597362816,
|
| 2530 |
+
"step": 970,
|
| 2531 |
+
"student_masked_tokens": 107.025
|
| 2532 |
+
},
|
| 2533 |
+
{
|
| 2534 |
+
"avg_mask_ratio": 0.42220073882490394,
|
| 2535 |
+
"avg_response_length": 230.8625,
|
| 2536 |
+
"avg_student_mask_ratio": 0.42220073882490394,
|
| 2537 |
+
"batch_ainp_frac": 0.0,
|
| 2538 |
+
"batch_inp_frac": 1.0,
|
| 2539 |
+
"batch_inp_oh_frac": 0.0,
|
| 2540 |
+
"batch_inp_par_frac": 0.0,
|
| 2541 |
+
"batch_inp_par_par_frac": 0.0,
|
| 2542 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 2543 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 2544 |
+
"batch_rl_frac": 0.0,
|
| 2545 |
+
"batch_sft_frac": 0.0,
|
| 2546 |
+
"batch_soft_sft_frac": 0.0,
|
| 2547 |
+
"batch_tf_frac": 0.0,
|
| 2548 |
+
"ce_loss": 0.0,
|
| 2549 |
+
"epoch": 2.0938666666666665,
|
| 2550 |
+
"grad_norm": 0.1943359375,
|
| 2551 |
+
"kd_loss": 0.3541594335687478,
|
| 2552 |
+
"learning_rate": 3e-06,
|
| 2553 |
+
"loss": 0.3645,
|
| 2554 |
+
"masked_tokens": 97.7875,
|
| 2555 |
+
"mean_t": 0.4511947895749472,
|
| 2556 |
+
"step": 980,
|
| 2557 |
+
"student_masked_tokens": 97.7875
|
| 2558 |
+
},
|
| 2559 |
+
{
|
| 2560 |
+
"avg_mask_ratio": 0.4605769342277199,
|
| 2561 |
+
"avg_response_length": 262.0375,
|
| 2562 |
+
"avg_student_mask_ratio": 0.4605769342277199,
|
| 2563 |
+
"batch_ainp_frac": 0.0,
|
| 2564 |
+
"batch_inp_frac": 1.0,
|
| 2565 |
+
"batch_inp_oh_frac": 0.0,
|
| 2566 |
+
"batch_inp_par_frac": 0.0,
|
| 2567 |
+
"batch_inp_par_par_frac": 0.0,
|
| 2568 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 2569 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 2570 |
+
"batch_rl_frac": 0.0,
|
| 2571 |
+
"batch_sft_frac": 0.0,
|
| 2572 |
+
"batch_soft_sft_frac": 0.0,
|
| 2573 |
+
"batch_tf_frac": 0.0,
|
| 2574 |
+
"ce_loss": 0.0,
|
| 2575 |
+
"epoch": 2.1152,
|
| 2576 |
+
"grad_norm": 0.2216796875,
|
| 2577 |
+
"kd_loss": 0.41152406695375704,
|
| 2578 |
+
"learning_rate": 3e-06,
|
| 2579 |
+
"loss": 0.4128,
|
| 2580 |
+
"masked_tokens": 122.0,
|
| 2581 |
+
"mean_t": 0.4923786667350214,
|
| 2582 |
+
"step": 990,
|
| 2583 |
+
"student_masked_tokens": 122.0
|
| 2584 |
+
},
|
| 2585 |
+
{
|
| 2586 |
+
"avg_mask_ratio": 0.4547682981239632,
|
| 2587 |
+
"avg_response_length": 215.3,
|
| 2588 |
+
"avg_student_mask_ratio": 0.4547682981239632,
|
| 2589 |
+
"batch_ainp_frac": 0.0,
|
| 2590 |
+
"batch_inp_frac": 1.0,
|
| 2591 |
+
"batch_inp_oh_frac": 0.0,
|
| 2592 |
+
"batch_inp_par_frac": 0.0,
|
| 2593 |
+
"batch_inp_par_par_frac": 0.0,
|
| 2594 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 2595 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 2596 |
+
"batch_rl_frac": 0.0,
|
| 2597 |
+
"batch_sft_frac": 0.0,
|
| 2598 |
+
"batch_soft_sft_frac": 0.0,
|
| 2599 |
+
"batch_tf_frac": 0.0,
|
| 2600 |
+
"ce_loss": 0.0,
|
| 2601 |
+
"epoch": 2.1365333333333334,
|
| 2602 |
+
"grad_norm": 0.326171875,
|
| 2603 |
+
"kd_loss": 0.3404828482017706,
|
| 2604 |
+
"learning_rate": 3e-06,
|
| 2605 |
+
"loss": 0.3793,
|
| 2606 |
+
"masked_tokens": 100.775,
|
| 2607 |
+
"mean_t": 0.4773523230338469,
|
| 2608 |
+
"step": 1000,
|
| 2609 |
+
"student_masked_tokens": 100.775
|
| 2610 |
+
}
|
| 2611 |
+
],
|
| 2612 |
+
"logging_steps": 10,
|
| 2613 |
+
"max_steps": 1404,
|
| 2614 |
+
"num_input_tokens_seen": 0,
|
| 2615 |
+
"num_train_epochs": 3,
|
| 2616 |
+
"save_steps": 100,
|
| 2617 |
+
"stateful_callbacks": {
|
| 2618 |
+
"TrainerControl": {
|
| 2619 |
+
"args": {
|
| 2620 |
+
"should_epoch_stop": false,
|
| 2621 |
+
"should_evaluate": false,
|
| 2622 |
+
"should_log": false,
|
| 2623 |
+
"should_save": true,
|
| 2624 |
+
"should_training_stop": false
|
| 2625 |
+
},
|
| 2626 |
+
"attributes": {}
|
| 2627 |
+
}
|
| 2628 |
+
},
|
| 2629 |
+
"total_flos": 0.0,
|
| 2630 |
+
"train_batch_size": 1,
|
| 2631 |
+
"trial_name": null,
|
| 2632 |
+
"trial_params": null
|
| 2633 |
+
}
|
math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-1000/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:99e6e32f3c28324ecee9c50eb13825addacaf6ce31b5b6707966fc1ac725dea3
|
| 3 |
+
size 8312
|
math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-1100/README.md
ADDED
|
@@ -0,0 +1,202 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
base_model: GSAI-ML/LLaDA-8B-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 0.15.1
|
math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-1100/adapter_config.json
ADDED
|
@@ -0,0 +1,39 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"alpha_pattern": {},
|
| 3 |
+
"auto_mapping": null,
|
| 4 |
+
"base_model_name_or_path": "GSAI-ML/LLaDA-8B-Instruct",
|
| 5 |
+
"bias": "none",
|
| 6 |
+
"corda_config": null,
|
| 7 |
+
"eva_config": null,
|
| 8 |
+
"exclude_modules": null,
|
| 9 |
+
"fan_in_fan_out": false,
|
| 10 |
+
"inference_mode": true,
|
| 11 |
+
"init_lora_weights": true,
|
| 12 |
+
"layer_replication": null,
|
| 13 |
+
"layers_pattern": null,
|
| 14 |
+
"layers_to_transform": null,
|
| 15 |
+
"loftq_config": {},
|
| 16 |
+
"lora_alpha": 64,
|
| 17 |
+
"lora_bias": false,
|
| 18 |
+
"lora_dropout": 0.05,
|
| 19 |
+
"megatron_config": null,
|
| 20 |
+
"megatron_core": "megatron.core",
|
| 21 |
+
"modules_to_save": null,
|
| 22 |
+
"peft_type": "LORA",
|
| 23 |
+
"r": 128,
|
| 24 |
+
"rank_pattern": {},
|
| 25 |
+
"revision": null,
|
| 26 |
+
"target_modules": [
|
| 27 |
+
"q_proj",
|
| 28 |
+
"k_proj",
|
| 29 |
+
"o_proj",
|
| 30 |
+
"up_proj",
|
| 31 |
+
"gate_proj",
|
| 32 |
+
"v_proj",
|
| 33 |
+
"down_proj"
|
| 34 |
+
],
|
| 35 |
+
"task_type": "CAUSAL_LM",
|
| 36 |
+
"trainable_token_indices": null,
|
| 37 |
+
"use_dora": false,
|
| 38 |
+
"use_rslora": false
|
| 39 |
+
}
|
math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-1100/adapter_model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5a3c53a6ecd3a30340175ad111a7fd1d734803f7005ed01db98ce7a014066274
|
| 3 |
+
size 2406624648
|
math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-1100/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b2a9bac61de5e3e842f2f3ce23538979dbb9e4f2bfa997b179b1651ec9976f0f
|
| 3 |
+
size 671304442
|
math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-1100/rng_state_0.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:907e39dff0cf7ad1a1affaa1e7047653794ab16e25c6977ce7b5524769fdf799
|
| 3 |
+
size 14512
|
math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-1100/rng_state_1.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ebf0b85ec66092ce471ce68f6fcda64182d50a012837c2e4c9f9690fe3e5f5c3
|
| 3 |
+
size 14512
|
math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-1100/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:25795e3b7374d0f6abdd7ab4b34fbf7ab0447ba73c04014500c2ab8b5acec5b4
|
| 3 |
+
size 1064
|
math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-1100/trainer_state.json
ADDED
|
@@ -0,0 +1,2893 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 2.3498666666666668,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 1100,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"avg_mask_ratio": 0.4931091487989761,
|
| 13 |
+
"avg_response_length": 225.975,
|
| 14 |
+
"avg_student_mask_ratio": 0.4931091487989761,
|
| 15 |
+
"batch_ainp_frac": 0.0,
|
| 16 |
+
"batch_inp_frac": 1.0,
|
| 17 |
+
"batch_inp_oh_frac": 0.0,
|
| 18 |
+
"batch_inp_par_frac": 0.0,
|
| 19 |
+
"batch_inp_par_par_frac": 0.0,
|
| 20 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 21 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 22 |
+
"batch_rl_frac": 0.0,
|
| 23 |
+
"batch_sft_frac": 0.0,
|
| 24 |
+
"batch_soft_sft_frac": 0.0,
|
| 25 |
+
"batch_tf_frac": 0.0,
|
| 26 |
+
"ce_loss": 0.0,
|
| 27 |
+
"epoch": 0.021333333333333333,
|
| 28 |
+
"grad_norm": 0.1953125,
|
| 29 |
+
"kd_loss": 0.537169519480176,
|
| 30 |
+
"learning_rate": 3e-06,
|
| 31 |
+
"loss": 0.5508,
|
| 32 |
+
"masked_tokens": 111.95,
|
| 33 |
+
"mean_t": 0.5145528071501758,
|
| 34 |
+
"step": 10,
|
| 35 |
+
"student_masked_tokens": 111.95
|
| 36 |
+
},
|
| 37 |
+
{
|
| 38 |
+
"avg_mask_ratio": 0.42058031369233506,
|
| 39 |
+
"avg_response_length": 255.2625,
|
| 40 |
+
"avg_student_mask_ratio": 0.42058031369233506,
|
| 41 |
+
"batch_ainp_frac": 0.0,
|
| 42 |
+
"batch_inp_frac": 1.0,
|
| 43 |
+
"batch_inp_oh_frac": 0.0,
|
| 44 |
+
"batch_inp_par_frac": 0.0,
|
| 45 |
+
"batch_inp_par_par_frac": 0.0,
|
| 46 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 47 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 48 |
+
"batch_rl_frac": 0.0,
|
| 49 |
+
"batch_sft_frac": 0.0,
|
| 50 |
+
"batch_soft_sft_frac": 0.0,
|
| 51 |
+
"batch_tf_frac": 0.0,
|
| 52 |
+
"ce_loss": 0.0,
|
| 53 |
+
"epoch": 0.042666666666666665,
|
| 54 |
+
"grad_norm": 0.23828125,
|
| 55 |
+
"kd_loss": 0.44732311753604304,
|
| 56 |
+
"learning_rate": 3e-06,
|
| 57 |
+
"loss": 0.3866,
|
| 58 |
+
"masked_tokens": 97.35,
|
| 59 |
+
"mean_t": 0.43874448732240123,
|
| 60 |
+
"step": 20,
|
| 61 |
+
"student_masked_tokens": 97.35
|
| 62 |
+
},
|
| 63 |
+
{
|
| 64 |
+
"avg_mask_ratio": 0.4538542575784959,
|
| 65 |
+
"avg_response_length": 211.7625,
|
| 66 |
+
"avg_student_mask_ratio": 0.4538542575784959,
|
| 67 |
+
"batch_ainp_frac": 0.0,
|
| 68 |
+
"batch_inp_frac": 1.0,
|
| 69 |
+
"batch_inp_oh_frac": 0.0,
|
| 70 |
+
"batch_inp_par_frac": 0.0,
|
| 71 |
+
"batch_inp_par_par_frac": 0.0,
|
| 72 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 73 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 74 |
+
"batch_rl_frac": 0.0,
|
| 75 |
+
"batch_sft_frac": 0.0,
|
| 76 |
+
"batch_soft_sft_frac": 0.0,
|
| 77 |
+
"batch_tf_frac": 0.0,
|
| 78 |
+
"ce_loss": 0.0,
|
| 79 |
+
"epoch": 0.064,
|
| 80 |
+
"grad_norm": 0.09130859375,
|
| 81 |
+
"kd_loss": 0.5295996250815808,
|
| 82 |
+
"learning_rate": 3e-06,
|
| 83 |
+
"loss": 0.462,
|
| 84 |
+
"masked_tokens": 110.2,
|
| 85 |
+
"mean_t": 0.4803953981841914,
|
| 86 |
+
"step": 30,
|
| 87 |
+
"student_masked_tokens": 110.2
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"avg_mask_ratio": 0.4207469140383182,
|
| 91 |
+
"avg_response_length": 224.125,
|
| 92 |
+
"avg_student_mask_ratio": 0.4207469140383182,
|
| 93 |
+
"batch_ainp_frac": 0.0,
|
| 94 |
+
"batch_inp_frac": 1.0,
|
| 95 |
+
"batch_inp_oh_frac": 0.0,
|
| 96 |
+
"batch_inp_par_frac": 0.0,
|
| 97 |
+
"batch_inp_par_par_frac": 0.0,
|
| 98 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 99 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 100 |
+
"batch_rl_frac": 0.0,
|
| 101 |
+
"batch_sft_frac": 0.0,
|
| 102 |
+
"batch_soft_sft_frac": 0.0,
|
| 103 |
+
"batch_tf_frac": 0.0,
|
| 104 |
+
"ce_loss": 0.0,
|
| 105 |
+
"epoch": 0.08533333333333333,
|
| 106 |
+
"grad_norm": 0.294921875,
|
| 107 |
+
"kd_loss": 0.3314848062681676,
|
| 108 |
+
"learning_rate": 3e-06,
|
| 109 |
+
"loss": 0.3521,
|
| 110 |
+
"masked_tokens": 98.1625,
|
| 111 |
+
"mean_t": 0.4569831106782658,
|
| 112 |
+
"step": 40,
|
| 113 |
+
"student_masked_tokens": 98.1625
|
| 114 |
+
},
|
| 115 |
+
{
|
| 116 |
+
"avg_mask_ratio": 0.43260439952719026,
|
| 117 |
+
"avg_response_length": 207.125,
|
| 118 |
+
"avg_student_mask_ratio": 0.43260439952719026,
|
| 119 |
+
"batch_ainp_frac": 0.0,
|
| 120 |
+
"batch_inp_frac": 1.0,
|
| 121 |
+
"batch_inp_oh_frac": 0.0,
|
| 122 |
+
"batch_inp_par_frac": 0.0,
|
| 123 |
+
"batch_inp_par_par_frac": 0.0,
|
| 124 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 125 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 126 |
+
"batch_rl_frac": 0.0,
|
| 127 |
+
"batch_sft_frac": 0.0,
|
| 128 |
+
"batch_soft_sft_frac": 0.0,
|
| 129 |
+
"batch_tf_frac": 0.0,
|
| 130 |
+
"ce_loss": 0.0,
|
| 131 |
+
"epoch": 0.10666666666666667,
|
| 132 |
+
"grad_norm": 0.1328125,
|
| 133 |
+
"kd_loss": 0.4259288739462136,
|
| 134 |
+
"learning_rate": 3e-06,
|
| 135 |
+
"loss": 0.4221,
|
| 136 |
+
"masked_tokens": 85.05,
|
| 137 |
+
"mean_t": 0.4612453707959503,
|
| 138 |
+
"step": 50,
|
| 139 |
+
"student_masked_tokens": 85.05
|
| 140 |
+
},
|
| 141 |
+
{
|
| 142 |
+
"avg_mask_ratio": 0.46053453313652426,
|
| 143 |
+
"avg_response_length": 251.0875,
|
| 144 |
+
"avg_student_mask_ratio": 0.46053453313652426,
|
| 145 |
+
"batch_ainp_frac": 0.0,
|
| 146 |
+
"batch_inp_frac": 1.0,
|
| 147 |
+
"batch_inp_oh_frac": 0.0,
|
| 148 |
+
"batch_inp_par_frac": 0.0,
|
| 149 |
+
"batch_inp_par_par_frac": 0.0,
|
| 150 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 151 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 152 |
+
"batch_rl_frac": 0.0,
|
| 153 |
+
"batch_sft_frac": 0.0,
|
| 154 |
+
"batch_soft_sft_frac": 0.0,
|
| 155 |
+
"batch_tf_frac": 0.0,
|
| 156 |
+
"ce_loss": 0.0,
|
| 157 |
+
"epoch": 0.128,
|
| 158 |
+
"grad_norm": 0.126953125,
|
| 159 |
+
"kd_loss": 0.38882399915036386,
|
| 160 |
+
"learning_rate": 3e-06,
|
| 161 |
+
"loss": 0.4142,
|
| 162 |
+
"masked_tokens": 120.9,
|
| 163 |
+
"mean_t": 0.48597636765334756,
|
| 164 |
+
"step": 60,
|
| 165 |
+
"student_masked_tokens": 120.9
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"avg_mask_ratio": 0.5112146578729153,
|
| 169 |
+
"avg_response_length": 202.5875,
|
| 170 |
+
"avg_student_mask_ratio": 0.5112146578729153,
|
| 171 |
+
"batch_ainp_frac": 0.0,
|
| 172 |
+
"batch_inp_frac": 1.0,
|
| 173 |
+
"batch_inp_oh_frac": 0.0,
|
| 174 |
+
"batch_inp_par_frac": 0.0,
|
| 175 |
+
"batch_inp_par_par_frac": 0.0,
|
| 176 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 177 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 178 |
+
"batch_rl_frac": 0.0,
|
| 179 |
+
"batch_sft_frac": 0.0,
|
| 180 |
+
"batch_soft_sft_frac": 0.0,
|
| 181 |
+
"batch_tf_frac": 0.0,
|
| 182 |
+
"ce_loss": 0.0,
|
| 183 |
+
"epoch": 0.14933333333333335,
|
| 184 |
+
"grad_norm": 0.3671875,
|
| 185 |
+
"kd_loss": 0.4423349894936791,
|
| 186 |
+
"learning_rate": 3e-06,
|
| 187 |
+
"loss": 0.4892,
|
| 188 |
+
"masked_tokens": 104.5875,
|
| 189 |
+
"mean_t": 0.5459650319069624,
|
| 190 |
+
"step": 70,
|
| 191 |
+
"student_masked_tokens": 104.5875
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"avg_mask_ratio": 0.37548826879356056,
|
| 195 |
+
"avg_response_length": 225.85,
|
| 196 |
+
"avg_student_mask_ratio": 0.37548826879356056,
|
| 197 |
+
"batch_ainp_frac": 0.0,
|
| 198 |
+
"batch_inp_frac": 1.0,
|
| 199 |
+
"batch_inp_oh_frac": 0.0,
|
| 200 |
+
"batch_inp_par_frac": 0.0,
|
| 201 |
+
"batch_inp_par_par_frac": 0.0,
|
| 202 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 203 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 204 |
+
"batch_rl_frac": 0.0,
|
| 205 |
+
"batch_sft_frac": 0.0,
|
| 206 |
+
"batch_soft_sft_frac": 0.0,
|
| 207 |
+
"batch_tf_frac": 0.0,
|
| 208 |
+
"ce_loss": 0.0,
|
| 209 |
+
"epoch": 0.17066666666666666,
|
| 210 |
+
"grad_norm": 0.083984375,
|
| 211 |
+
"kd_loss": 0.3107750686068783,
|
| 212 |
+
"learning_rate": 3e-06,
|
| 213 |
+
"loss": 0.3183,
|
| 214 |
+
"masked_tokens": 85.0625,
|
| 215 |
+
"mean_t": 0.40758824030635876,
|
| 216 |
+
"step": 80,
|
| 217 |
+
"student_masked_tokens": 85.0625
|
| 218 |
+
},
|
| 219 |
+
{
|
| 220 |
+
"avg_mask_ratio": 0.5001560213277116,
|
| 221 |
+
"avg_response_length": 229.75,
|
| 222 |
+
"avg_student_mask_ratio": 0.5001560213277116,
|
| 223 |
+
"batch_ainp_frac": 0.0,
|
| 224 |
+
"batch_inp_frac": 1.0,
|
| 225 |
+
"batch_inp_oh_frac": 0.0,
|
| 226 |
+
"batch_inp_par_frac": 0.0,
|
| 227 |
+
"batch_inp_par_par_frac": 0.0,
|
| 228 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 229 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 230 |
+
"batch_rl_frac": 0.0,
|
| 231 |
+
"batch_sft_frac": 0.0,
|
| 232 |
+
"batch_soft_sft_frac": 0.0,
|
| 233 |
+
"batch_tf_frac": 0.0,
|
| 234 |
+
"ce_loss": 0.0,
|
| 235 |
+
"epoch": 0.192,
|
| 236 |
+
"grad_norm": 0.57421875,
|
| 237 |
+
"kd_loss": 0.600051444793823,
|
| 238 |
+
"learning_rate": 3e-06,
|
| 239 |
+
"loss": 0.578,
|
| 240 |
+
"masked_tokens": 107.6625,
|
| 241 |
+
"mean_t": 0.5297661645396147,
|
| 242 |
+
"step": 90,
|
| 243 |
+
"student_masked_tokens": 107.6625
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"avg_mask_ratio": 0.4576045103633078,
|
| 247 |
+
"avg_response_length": 208.0,
|
| 248 |
+
"avg_student_mask_ratio": 0.4576045103633078,
|
| 249 |
+
"batch_ainp_frac": 0.0,
|
| 250 |
+
"batch_inp_frac": 1.0,
|
| 251 |
+
"batch_inp_oh_frac": 0.0,
|
| 252 |
+
"batch_inp_par_frac": 0.0,
|
| 253 |
+
"batch_inp_par_par_frac": 0.0,
|
| 254 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 255 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 256 |
+
"batch_rl_frac": 0.0,
|
| 257 |
+
"batch_sft_frac": 0.0,
|
| 258 |
+
"batch_soft_sft_frac": 0.0,
|
| 259 |
+
"batch_tf_frac": 0.0,
|
| 260 |
+
"ce_loss": 0.0,
|
| 261 |
+
"epoch": 0.21333333333333335,
|
| 262 |
+
"grad_norm": 0.1767578125,
|
| 263 |
+
"kd_loss": 0.38177184401774583,
|
| 264 |
+
"learning_rate": 3e-06,
|
| 265 |
+
"loss": 0.4107,
|
| 266 |
+
"masked_tokens": 104.1625,
|
| 267 |
+
"mean_t": 0.4886587227345444,
|
| 268 |
+
"step": 100,
|
| 269 |
+
"student_masked_tokens": 104.1625
|
| 270 |
+
},
|
| 271 |
+
{
|
| 272 |
+
"avg_mask_ratio": 0.4877026333590038,
|
| 273 |
+
"avg_response_length": 213.0875,
|
| 274 |
+
"avg_student_mask_ratio": 0.4877026333590038,
|
| 275 |
+
"batch_ainp_frac": 0.0,
|
| 276 |
+
"batch_inp_frac": 1.0,
|
| 277 |
+
"batch_inp_oh_frac": 0.0,
|
| 278 |
+
"batch_inp_par_frac": 0.0,
|
| 279 |
+
"batch_inp_par_par_frac": 0.0,
|
| 280 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 281 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 282 |
+
"batch_rl_frac": 0.0,
|
| 283 |
+
"batch_sft_frac": 0.0,
|
| 284 |
+
"batch_soft_sft_frac": 0.0,
|
| 285 |
+
"batch_tf_frac": 0.0,
|
| 286 |
+
"ce_loss": 0.0,
|
| 287 |
+
"epoch": 0.23466666666666666,
|
| 288 |
+
"grad_norm": 0.333984375,
|
| 289 |
+
"kd_loss": 0.50661907401431,
|
| 290 |
+
"learning_rate": 3e-06,
|
| 291 |
+
"loss": 0.5293,
|
| 292 |
+
"masked_tokens": 102.075,
|
| 293 |
+
"mean_t": 0.5246987929102034,
|
| 294 |
+
"step": 110,
|
| 295 |
+
"student_masked_tokens": 102.075
|
| 296 |
+
},
|
| 297 |
+
{
|
| 298 |
+
"avg_mask_ratio": 0.45146879020612685,
|
| 299 |
+
"avg_response_length": 224.1875,
|
| 300 |
+
"avg_student_mask_ratio": 0.45146879020612685,
|
| 301 |
+
"batch_ainp_frac": 0.0,
|
| 302 |
+
"batch_inp_frac": 1.0,
|
| 303 |
+
"batch_inp_oh_frac": 0.0,
|
| 304 |
+
"batch_inp_par_frac": 0.0,
|
| 305 |
+
"batch_inp_par_par_frac": 0.0,
|
| 306 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 307 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 308 |
+
"batch_rl_frac": 0.0,
|
| 309 |
+
"batch_sft_frac": 0.0,
|
| 310 |
+
"batch_soft_sft_frac": 0.0,
|
| 311 |
+
"batch_tf_frac": 0.0,
|
| 312 |
+
"ce_loss": 0.0,
|
| 313 |
+
"epoch": 0.256,
|
| 314 |
+
"grad_norm": 0.236328125,
|
| 315 |
+
"kd_loss": 0.4163589179806877,
|
| 316 |
+
"learning_rate": 3e-06,
|
| 317 |
+
"loss": 0.435,
|
| 318 |
+
"masked_tokens": 100.525,
|
| 319 |
+
"mean_t": 0.4805434140143916,
|
| 320 |
+
"step": 120,
|
| 321 |
+
"student_masked_tokens": 100.525
|
| 322 |
+
},
|
| 323 |
+
{
|
| 324 |
+
"avg_mask_ratio": 0.4356566035945434,
|
| 325 |
+
"avg_response_length": 202.7,
|
| 326 |
+
"avg_student_mask_ratio": 0.4356566035945434,
|
| 327 |
+
"batch_ainp_frac": 0.0,
|
| 328 |
+
"batch_inp_frac": 1.0,
|
| 329 |
+
"batch_inp_oh_frac": 0.0,
|
| 330 |
+
"batch_inp_par_frac": 0.0,
|
| 331 |
+
"batch_inp_par_par_frac": 0.0,
|
| 332 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 333 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 334 |
+
"batch_rl_frac": 0.0,
|
| 335 |
+
"batch_sft_frac": 0.0,
|
| 336 |
+
"batch_soft_sft_frac": 0.0,
|
| 337 |
+
"batch_tf_frac": 0.0,
|
| 338 |
+
"ce_loss": 0.0,
|
| 339 |
+
"epoch": 0.2773333333333333,
|
| 340 |
+
"grad_norm": 0.2314453125,
|
| 341 |
+
"kd_loss": 0.2308679626777348,
|
| 342 |
+
"learning_rate": 3e-06,
|
| 343 |
+
"loss": 0.2829,
|
| 344 |
+
"masked_tokens": 89.5625,
|
| 345 |
+
"mean_t": 0.47522516988683494,
|
| 346 |
+
"step": 130,
|
| 347 |
+
"student_masked_tokens": 89.5625
|
| 348 |
+
},
|
| 349 |
+
{
|
| 350 |
+
"avg_mask_ratio": 0.49419954856857656,
|
| 351 |
+
"avg_response_length": 255.625,
|
| 352 |
+
"avg_student_mask_ratio": 0.49419954856857656,
|
| 353 |
+
"batch_ainp_frac": 0.0,
|
| 354 |
+
"batch_inp_frac": 1.0,
|
| 355 |
+
"batch_inp_oh_frac": 0.0,
|
| 356 |
+
"batch_inp_par_frac": 0.0,
|
| 357 |
+
"batch_inp_par_par_frac": 0.0,
|
| 358 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 359 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 360 |
+
"batch_rl_frac": 0.0,
|
| 361 |
+
"batch_sft_frac": 0.0,
|
| 362 |
+
"batch_soft_sft_frac": 0.0,
|
| 363 |
+
"batch_tf_frac": 0.0,
|
| 364 |
+
"ce_loss": 0.0,
|
| 365 |
+
"epoch": 0.2986666666666667,
|
| 366 |
+
"grad_norm": 0.353515625,
|
| 367 |
+
"kd_loss": 0.5027470885725507,
|
| 368 |
+
"learning_rate": 3e-06,
|
| 369 |
+
"loss": 0.5364,
|
| 370 |
+
"masked_tokens": 136.575,
|
| 371 |
+
"mean_t": 0.5204090005659964,
|
| 372 |
+
"step": 140,
|
| 373 |
+
"student_masked_tokens": 136.575
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"avg_mask_ratio": 0.4736677930341102,
|
| 377 |
+
"avg_response_length": 255.375,
|
| 378 |
+
"avg_student_mask_ratio": 0.4736677930341102,
|
| 379 |
+
"batch_ainp_frac": 0.0,
|
| 380 |
+
"batch_inp_frac": 1.0,
|
| 381 |
+
"batch_inp_oh_frac": 0.0,
|
| 382 |
+
"batch_inp_par_frac": 0.0,
|
| 383 |
+
"batch_inp_par_par_frac": 0.0,
|
| 384 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 385 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 386 |
+
"batch_rl_frac": 0.0,
|
| 387 |
+
"batch_sft_frac": 0.0,
|
| 388 |
+
"batch_soft_sft_frac": 0.0,
|
| 389 |
+
"batch_tf_frac": 0.0,
|
| 390 |
+
"ce_loss": 0.0,
|
| 391 |
+
"epoch": 0.32,
|
| 392 |
+
"grad_norm": 0.07666015625,
|
| 393 |
+
"kd_loss": 0.4891943664209066,
|
| 394 |
+
"learning_rate": 3e-06,
|
| 395 |
+
"loss": 0.5326,
|
| 396 |
+
"masked_tokens": 126.575,
|
| 397 |
+
"mean_t": 0.5012552456930279,
|
| 398 |
+
"step": 150,
|
| 399 |
+
"student_masked_tokens": 126.575
|
| 400 |
+
},
|
| 401 |
+
{
|
| 402 |
+
"avg_mask_ratio": 0.5032523009285796,
|
| 403 |
+
"avg_response_length": 209.325,
|
| 404 |
+
"avg_student_mask_ratio": 0.5032523009285796,
|
| 405 |
+
"batch_ainp_frac": 0.0,
|
| 406 |
+
"batch_inp_frac": 1.0,
|
| 407 |
+
"batch_inp_oh_frac": 0.0,
|
| 408 |
+
"batch_inp_par_frac": 0.0,
|
| 409 |
+
"batch_inp_par_par_frac": 0.0,
|
| 410 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 411 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 412 |
+
"batch_rl_frac": 0.0,
|
| 413 |
+
"batch_sft_frac": 0.0,
|
| 414 |
+
"batch_soft_sft_frac": 0.0,
|
| 415 |
+
"batch_tf_frac": 0.0,
|
| 416 |
+
"ce_loss": 0.0,
|
| 417 |
+
"epoch": 0.3413333333333333,
|
| 418 |
+
"grad_norm": 0.353515625,
|
| 419 |
+
"kd_loss": 0.4466663982801265,
|
| 420 |
+
"learning_rate": 3e-06,
|
| 421 |
+
"loss": 0.487,
|
| 422 |
+
"masked_tokens": 99.425,
|
| 423 |
+
"mean_t": 0.5408745193795766,
|
| 424 |
+
"step": 160,
|
| 425 |
+
"student_masked_tokens": 99.425
|
| 426 |
+
},
|
| 427 |
+
{
|
| 428 |
+
"avg_mask_ratio": 0.4806730231270194,
|
| 429 |
+
"avg_response_length": 190.325,
|
| 430 |
+
"avg_student_mask_ratio": 0.4806730231270194,
|
| 431 |
+
"batch_ainp_frac": 0.0,
|
| 432 |
+
"batch_inp_frac": 1.0,
|
| 433 |
+
"batch_inp_oh_frac": 0.0,
|
| 434 |
+
"batch_inp_par_frac": 0.0,
|
| 435 |
+
"batch_inp_par_par_frac": 0.0,
|
| 436 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 437 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 438 |
+
"batch_rl_frac": 0.0,
|
| 439 |
+
"batch_sft_frac": 0.0,
|
| 440 |
+
"batch_soft_sft_frac": 0.0,
|
| 441 |
+
"batch_tf_frac": 0.0,
|
| 442 |
+
"ce_loss": 0.0,
|
| 443 |
+
"epoch": 0.3626666666666667,
|
| 444 |
+
"grad_norm": 0.1650390625,
|
| 445 |
+
"kd_loss": 0.45357214291389025,
|
| 446 |
+
"learning_rate": 3e-06,
|
| 447 |
+
"loss": 0.4277,
|
| 448 |
+
"masked_tokens": 89.825,
|
| 449 |
+
"mean_t": 0.5134547733236104,
|
| 450 |
+
"step": 170,
|
| 451 |
+
"student_masked_tokens": 89.825
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"avg_mask_ratio": 0.45829249716189224,
|
| 455 |
+
"avg_response_length": 244.0,
|
| 456 |
+
"avg_student_mask_ratio": 0.45829249716189224,
|
| 457 |
+
"batch_ainp_frac": 0.0,
|
| 458 |
+
"batch_inp_frac": 1.0,
|
| 459 |
+
"batch_inp_oh_frac": 0.0,
|
| 460 |
+
"batch_inp_par_frac": 0.0,
|
| 461 |
+
"batch_inp_par_par_frac": 0.0,
|
| 462 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 463 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 464 |
+
"batch_rl_frac": 0.0,
|
| 465 |
+
"batch_sft_frac": 0.0,
|
| 466 |
+
"batch_soft_sft_frac": 0.0,
|
| 467 |
+
"batch_tf_frac": 0.0,
|
| 468 |
+
"ce_loss": 0.0,
|
| 469 |
+
"epoch": 0.384,
|
| 470 |
+
"grad_norm": 0.205078125,
|
| 471 |
+
"kd_loss": 0.35104387890574457,
|
| 472 |
+
"learning_rate": 3e-06,
|
| 473 |
+
"loss": 0.3341,
|
| 474 |
+
"masked_tokens": 110.075,
|
| 475 |
+
"mean_t": 0.48226988823735156,
|
| 476 |
+
"step": 180,
|
| 477 |
+
"student_masked_tokens": 110.075
|
| 478 |
+
},
|
| 479 |
+
{
|
| 480 |
+
"avg_mask_ratio": 0.4842760307248682,
|
| 481 |
+
"avg_response_length": 233.675,
|
| 482 |
+
"avg_student_mask_ratio": 0.4842760307248682,
|
| 483 |
+
"batch_ainp_frac": 0.0,
|
| 484 |
+
"batch_inp_frac": 1.0,
|
| 485 |
+
"batch_inp_oh_frac": 0.0,
|
| 486 |
+
"batch_inp_par_frac": 0.0,
|
| 487 |
+
"batch_inp_par_par_frac": 0.0,
|
| 488 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 489 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 490 |
+
"batch_rl_frac": 0.0,
|
| 491 |
+
"batch_sft_frac": 0.0,
|
| 492 |
+
"batch_soft_sft_frac": 0.0,
|
| 493 |
+
"batch_tf_frac": 0.0,
|
| 494 |
+
"ce_loss": 0.0,
|
| 495 |
+
"epoch": 0.4053333333333333,
|
| 496 |
+
"grad_norm": 0.84375,
|
| 497 |
+
"kd_loss": 0.43071598413393986,
|
| 498 |
+
"learning_rate": 3e-06,
|
| 499 |
+
"loss": 0.4677,
|
| 500 |
+
"masked_tokens": 109.5875,
|
| 501 |
+
"mean_t": 0.5165087037021294,
|
| 502 |
+
"step": 190,
|
| 503 |
+
"student_masked_tokens": 109.5875
|
| 504 |
+
},
|
| 505 |
+
{
|
| 506 |
+
"avg_mask_ratio": 0.4665210062637925,
|
| 507 |
+
"avg_response_length": 197.8,
|
| 508 |
+
"avg_student_mask_ratio": 0.4665210062637925,
|
| 509 |
+
"batch_ainp_frac": 0.0,
|
| 510 |
+
"batch_inp_frac": 1.0,
|
| 511 |
+
"batch_inp_oh_frac": 0.0,
|
| 512 |
+
"batch_inp_par_frac": 0.0,
|
| 513 |
+
"batch_inp_par_par_frac": 0.0,
|
| 514 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 515 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 516 |
+
"batch_rl_frac": 0.0,
|
| 517 |
+
"batch_sft_frac": 0.0,
|
| 518 |
+
"batch_soft_sft_frac": 0.0,
|
| 519 |
+
"batch_tf_frac": 0.0,
|
| 520 |
+
"ce_loss": 0.0,
|
| 521 |
+
"epoch": 0.4266666666666667,
|
| 522 |
+
"grad_norm": 0.2421875,
|
| 523 |
+
"kd_loss": 0.35780135999341384,
|
| 524 |
+
"learning_rate": 3e-06,
|
| 525 |
+
"loss": 0.3577,
|
| 526 |
+
"masked_tokens": 97.0125,
|
| 527 |
+
"mean_t": 0.5073627714533359,
|
| 528 |
+
"step": 200,
|
| 529 |
+
"student_masked_tokens": 97.0125
|
| 530 |
+
},
|
| 531 |
+
{
|
| 532 |
+
"avg_mask_ratio": 0.45183838629163803,
|
| 533 |
+
"avg_response_length": 213.7875,
|
| 534 |
+
"avg_student_mask_ratio": 0.45183838629163803,
|
| 535 |
+
"batch_ainp_frac": 0.0,
|
| 536 |
+
"batch_inp_frac": 1.0,
|
| 537 |
+
"batch_inp_oh_frac": 0.0,
|
| 538 |
+
"batch_inp_par_frac": 0.0,
|
| 539 |
+
"batch_inp_par_par_frac": 0.0,
|
| 540 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 541 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 542 |
+
"batch_rl_frac": 0.0,
|
| 543 |
+
"batch_sft_frac": 0.0,
|
| 544 |
+
"batch_soft_sft_frac": 0.0,
|
| 545 |
+
"batch_tf_frac": 0.0,
|
| 546 |
+
"ce_loss": 0.0,
|
| 547 |
+
"epoch": 0.448,
|
| 548 |
+
"grad_norm": 0.2138671875,
|
| 549 |
+
"kd_loss": 0.38703344948080487,
|
| 550 |
+
"learning_rate": 3e-06,
|
| 551 |
+
"loss": 0.3627,
|
| 552 |
+
"masked_tokens": 96.0125,
|
| 553 |
+
"mean_t": 0.48430291628465055,
|
| 554 |
+
"step": 210,
|
| 555 |
+
"student_masked_tokens": 96.0125
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"avg_mask_ratio": 0.5396152360364794,
|
| 559 |
+
"avg_response_length": 236.5125,
|
| 560 |
+
"avg_student_mask_ratio": 0.5396152360364794,
|
| 561 |
+
"batch_ainp_frac": 0.0,
|
| 562 |
+
"batch_inp_frac": 1.0,
|
| 563 |
+
"batch_inp_oh_frac": 0.0,
|
| 564 |
+
"batch_inp_par_frac": 0.0,
|
| 565 |
+
"batch_inp_par_par_frac": 0.0,
|
| 566 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 567 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 568 |
+
"batch_rl_frac": 0.0,
|
| 569 |
+
"batch_sft_frac": 0.0,
|
| 570 |
+
"batch_soft_sft_frac": 0.0,
|
| 571 |
+
"batch_tf_frac": 0.0,
|
| 572 |
+
"ce_loss": 0.0,
|
| 573 |
+
"epoch": 0.4693333333333333,
|
| 574 |
+
"grad_norm": 0.3203125,
|
| 575 |
+
"kd_loss": 0.4665673110896023,
|
| 576 |
+
"learning_rate": 3e-06,
|
| 577 |
+
"loss": 0.5258,
|
| 578 |
+
"masked_tokens": 125.35,
|
| 579 |
+
"mean_t": 0.570199209311977,
|
| 580 |
+
"step": 220,
|
| 581 |
+
"student_masked_tokens": 125.35
|
| 582 |
+
},
|
| 583 |
+
{
|
| 584 |
+
"avg_mask_ratio": 0.4403991688624956,
|
| 585 |
+
"avg_response_length": 252.7,
|
| 586 |
+
"avg_student_mask_ratio": 0.4403991688624956,
|
| 587 |
+
"batch_ainp_frac": 0.0,
|
| 588 |
+
"batch_inp_frac": 1.0,
|
| 589 |
+
"batch_inp_oh_frac": 0.0,
|
| 590 |
+
"batch_inp_par_frac": 0.0,
|
| 591 |
+
"batch_inp_par_par_frac": 0.0,
|
| 592 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 593 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 594 |
+
"batch_rl_frac": 0.0,
|
| 595 |
+
"batch_sft_frac": 0.0,
|
| 596 |
+
"batch_soft_sft_frac": 0.0,
|
| 597 |
+
"batch_tf_frac": 0.0,
|
| 598 |
+
"ce_loss": 0.0,
|
| 599 |
+
"epoch": 0.49066666666666664,
|
| 600 |
+
"grad_norm": 0.220703125,
|
| 601 |
+
"kd_loss": 0.4347595912117505,
|
| 602 |
+
"learning_rate": 3e-06,
|
| 603 |
+
"loss": 0.4555,
|
| 604 |
+
"masked_tokens": 107.325,
|
| 605 |
+
"mean_t": 0.46891279935371133,
|
| 606 |
+
"step": 230,
|
| 607 |
+
"student_masked_tokens": 107.325
|
| 608 |
+
},
|
| 609 |
+
{
|
| 610 |
+
"avg_mask_ratio": 0.47419991258066146,
|
| 611 |
+
"avg_response_length": 212.85,
|
| 612 |
+
"avg_student_mask_ratio": 0.47419991258066146,
|
| 613 |
+
"batch_ainp_frac": 0.0,
|
| 614 |
+
"batch_inp_frac": 1.0,
|
| 615 |
+
"batch_inp_oh_frac": 0.0,
|
| 616 |
+
"batch_inp_par_frac": 0.0,
|
| 617 |
+
"batch_inp_par_par_frac": 0.0,
|
| 618 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 619 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 620 |
+
"batch_rl_frac": 0.0,
|
| 621 |
+
"batch_sft_frac": 0.0,
|
| 622 |
+
"batch_soft_sft_frac": 0.0,
|
| 623 |
+
"batch_tf_frac": 0.0,
|
| 624 |
+
"ce_loss": 0.0,
|
| 625 |
+
"epoch": 0.512,
|
| 626 |
+
"grad_norm": 0.1044921875,
|
| 627 |
+
"kd_loss": 0.5054432464686386,
|
| 628 |
+
"learning_rate": 3e-06,
|
| 629 |
+
"loss": 0.5307,
|
| 630 |
+
"masked_tokens": 106.9,
|
| 631 |
+
"mean_t": 0.502228345896583,
|
| 632 |
+
"step": 240,
|
| 633 |
+
"student_masked_tokens": 106.9
|
| 634 |
+
},
|
| 635 |
+
{
|
| 636 |
+
"avg_mask_ratio": 0.4464209079160355,
|
| 637 |
+
"avg_response_length": 243.475,
|
| 638 |
+
"avg_student_mask_ratio": 0.4464209079160355,
|
| 639 |
+
"batch_ainp_frac": 0.0,
|
| 640 |
+
"batch_inp_frac": 1.0,
|
| 641 |
+
"batch_inp_oh_frac": 0.0,
|
| 642 |
+
"batch_inp_par_frac": 0.0,
|
| 643 |
+
"batch_inp_par_par_frac": 0.0,
|
| 644 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 645 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 646 |
+
"batch_rl_frac": 0.0,
|
| 647 |
+
"batch_sft_frac": 0.0,
|
| 648 |
+
"batch_soft_sft_frac": 0.0,
|
| 649 |
+
"batch_tf_frac": 0.0,
|
| 650 |
+
"ce_loss": 0.0,
|
| 651 |
+
"epoch": 0.5333333333333333,
|
| 652 |
+
"grad_norm": 0.1025390625,
|
| 653 |
+
"kd_loss": 0.41803101517930374,
|
| 654 |
+
"learning_rate": 3e-06,
|
| 655 |
+
"loss": 0.3942,
|
| 656 |
+
"masked_tokens": 112.7375,
|
| 657 |
+
"mean_t": 0.4733429416548461,
|
| 658 |
+
"step": 250,
|
| 659 |
+
"student_masked_tokens": 112.7375
|
| 660 |
+
},
|
| 661 |
+
{
|
| 662 |
+
"avg_mask_ratio": 0.4520751796895638,
|
| 663 |
+
"avg_response_length": 245.55,
|
| 664 |
+
"avg_student_mask_ratio": 0.4520751796895638,
|
| 665 |
+
"batch_ainp_frac": 0.0,
|
| 666 |
+
"batch_inp_frac": 1.0,
|
| 667 |
+
"batch_inp_oh_frac": 0.0,
|
| 668 |
+
"batch_inp_par_frac": 0.0,
|
| 669 |
+
"batch_inp_par_par_frac": 0.0,
|
| 670 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 671 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 672 |
+
"batch_rl_frac": 0.0,
|
| 673 |
+
"batch_sft_frac": 0.0,
|
| 674 |
+
"batch_soft_sft_frac": 0.0,
|
| 675 |
+
"batch_tf_frac": 0.0,
|
| 676 |
+
"ce_loss": 0.0,
|
| 677 |
+
"epoch": 0.5546666666666666,
|
| 678 |
+
"grad_norm": 0.1806640625,
|
| 679 |
+
"kd_loss": 0.3202606088300026,
|
| 680 |
+
"learning_rate": 3e-06,
|
| 681 |
+
"loss": 0.3513,
|
| 682 |
+
"masked_tokens": 109.6375,
|
| 683 |
+
"mean_t": 0.4843149524240289,
|
| 684 |
+
"step": 260,
|
| 685 |
+
"student_masked_tokens": 109.6375
|
| 686 |
+
},
|
| 687 |
+
{
|
| 688 |
+
"avg_mask_ratio": 0.5305180630879477,
|
| 689 |
+
"avg_response_length": 224.45,
|
| 690 |
+
"avg_student_mask_ratio": 0.5305180630879477,
|
| 691 |
+
"batch_ainp_frac": 0.0,
|
| 692 |
+
"batch_inp_frac": 1.0,
|
| 693 |
+
"batch_inp_oh_frac": 0.0,
|
| 694 |
+
"batch_inp_par_frac": 0.0,
|
| 695 |
+
"batch_inp_par_par_frac": 0.0,
|
| 696 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 697 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 698 |
+
"batch_rl_frac": 0.0,
|
| 699 |
+
"batch_sft_frac": 0.0,
|
| 700 |
+
"batch_soft_sft_frac": 0.0,
|
| 701 |
+
"batch_tf_frac": 0.0,
|
| 702 |
+
"ce_loss": 0.0,
|
| 703 |
+
"epoch": 0.576,
|
| 704 |
+
"grad_norm": 0.28515625,
|
| 705 |
+
"kd_loss": 0.5523448550798549,
|
| 706 |
+
"learning_rate": 3e-06,
|
| 707 |
+
"loss": 0.5602,
|
| 708 |
+
"masked_tokens": 120.475,
|
| 709 |
+
"mean_t": 0.5643589949700981,
|
| 710 |
+
"step": 270,
|
| 711 |
+
"student_masked_tokens": 120.475
|
| 712 |
+
},
|
| 713 |
+
{
|
| 714 |
+
"avg_mask_ratio": 0.46451686368091033,
|
| 715 |
+
"avg_response_length": 254.825,
|
| 716 |
+
"avg_student_mask_ratio": 0.46451686368091033,
|
| 717 |
+
"batch_ainp_frac": 0.0,
|
| 718 |
+
"batch_inp_frac": 1.0,
|
| 719 |
+
"batch_inp_oh_frac": 0.0,
|
| 720 |
+
"batch_inp_par_frac": 0.0,
|
| 721 |
+
"batch_inp_par_par_frac": 0.0,
|
| 722 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 723 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 724 |
+
"batch_rl_frac": 0.0,
|
| 725 |
+
"batch_sft_frac": 0.0,
|
| 726 |
+
"batch_soft_sft_frac": 0.0,
|
| 727 |
+
"batch_tf_frac": 0.0,
|
| 728 |
+
"ce_loss": 0.0,
|
| 729 |
+
"epoch": 0.5973333333333334,
|
| 730 |
+
"grad_norm": 0.2451171875,
|
| 731 |
+
"kd_loss": 0.3936164884111079,
|
| 732 |
+
"learning_rate": 3e-06,
|
| 733 |
+
"loss": 0.4356,
|
| 734 |
+
"masked_tokens": 129.225,
|
| 735 |
+
"mean_t": 0.47818811538163575,
|
| 736 |
+
"step": 280,
|
| 737 |
+
"student_masked_tokens": 129.225
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"avg_mask_ratio": 0.389662017847877,
|
| 741 |
+
"avg_response_length": 245.9125,
|
| 742 |
+
"avg_student_mask_ratio": 0.389662017847877,
|
| 743 |
+
"batch_ainp_frac": 0.0,
|
| 744 |
+
"batch_inp_frac": 1.0,
|
| 745 |
+
"batch_inp_oh_frac": 0.0,
|
| 746 |
+
"batch_inp_par_frac": 0.0,
|
| 747 |
+
"batch_inp_par_par_frac": 0.0,
|
| 748 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 749 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 750 |
+
"batch_rl_frac": 0.0,
|
| 751 |
+
"batch_sft_frac": 0.0,
|
| 752 |
+
"batch_soft_sft_frac": 0.0,
|
| 753 |
+
"batch_tf_frac": 0.0,
|
| 754 |
+
"ce_loss": 0.0,
|
| 755 |
+
"epoch": 0.6186666666666667,
|
| 756 |
+
"grad_norm": 0.10888671875,
|
| 757 |
+
"kd_loss": 0.27756736385699055,
|
| 758 |
+
"learning_rate": 3e-06,
|
| 759 |
+
"loss": 0.3305,
|
| 760 |
+
"masked_tokens": 99.8625,
|
| 761 |
+
"mean_t": 0.4088635521940887,
|
| 762 |
+
"step": 290,
|
| 763 |
+
"student_masked_tokens": 99.8625
|
| 764 |
+
},
|
| 765 |
+
{
|
| 766 |
+
"avg_mask_ratio": 0.44417000194080175,
|
| 767 |
+
"avg_response_length": 217.0375,
|
| 768 |
+
"avg_student_mask_ratio": 0.44417000194080175,
|
| 769 |
+
"batch_ainp_frac": 0.0,
|
| 770 |
+
"batch_inp_frac": 1.0,
|
| 771 |
+
"batch_inp_oh_frac": 0.0,
|
| 772 |
+
"batch_inp_par_frac": 0.0,
|
| 773 |
+
"batch_inp_par_par_frac": 0.0,
|
| 774 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 775 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 776 |
+
"batch_rl_frac": 0.0,
|
| 777 |
+
"batch_sft_frac": 0.0,
|
| 778 |
+
"batch_soft_sft_frac": 0.0,
|
| 779 |
+
"batch_tf_frac": 0.0,
|
| 780 |
+
"ce_loss": 0.0,
|
| 781 |
+
"epoch": 0.64,
|
| 782 |
+
"grad_norm": 0.16015625,
|
| 783 |
+
"kd_loss": 0.31748922848000805,
|
| 784 |
+
"learning_rate": 3e-06,
|
| 785 |
+
"loss": 0.3705,
|
| 786 |
+
"masked_tokens": 104.625,
|
| 787 |
+
"mean_t": 0.47984200695063917,
|
| 788 |
+
"step": 300,
|
| 789 |
+
"student_masked_tokens": 104.625
|
| 790 |
+
},
|
| 791 |
+
{
|
| 792 |
+
"avg_mask_ratio": 0.4706685543409549,
|
| 793 |
+
"avg_response_length": 175.45,
|
| 794 |
+
"avg_student_mask_ratio": 0.4706685543409549,
|
| 795 |
+
"batch_ainp_frac": 0.0,
|
| 796 |
+
"batch_inp_frac": 1.0,
|
| 797 |
+
"batch_inp_oh_frac": 0.0,
|
| 798 |
+
"batch_inp_par_frac": 0.0,
|
| 799 |
+
"batch_inp_par_par_frac": 0.0,
|
| 800 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 801 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 802 |
+
"batch_rl_frac": 0.0,
|
| 803 |
+
"batch_sft_frac": 0.0,
|
| 804 |
+
"batch_soft_sft_frac": 0.0,
|
| 805 |
+
"batch_tf_frac": 0.0,
|
| 806 |
+
"ce_loss": 0.0,
|
| 807 |
+
"epoch": 0.6613333333333333,
|
| 808 |
+
"grad_norm": 0.263671875,
|
| 809 |
+
"kd_loss": 0.5043777307220239,
|
| 810 |
+
"learning_rate": 3e-06,
|
| 811 |
+
"loss": 0.4515,
|
| 812 |
+
"masked_tokens": 84.875,
|
| 813 |
+
"mean_t": 0.5026606284547597,
|
| 814 |
+
"step": 310,
|
| 815 |
+
"student_masked_tokens": 84.875
|
| 816 |
+
},
|
| 817 |
+
{
|
| 818 |
+
"avg_mask_ratio": 0.4974605386145413,
|
| 819 |
+
"avg_response_length": 234.7875,
|
| 820 |
+
"avg_student_mask_ratio": 0.4974605386145413,
|
| 821 |
+
"batch_ainp_frac": 0.0,
|
| 822 |
+
"batch_inp_frac": 1.0,
|
| 823 |
+
"batch_inp_oh_frac": 0.0,
|
| 824 |
+
"batch_inp_par_frac": 0.0,
|
| 825 |
+
"batch_inp_par_par_frac": 0.0,
|
| 826 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 827 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 828 |
+
"batch_rl_frac": 0.0,
|
| 829 |
+
"batch_sft_frac": 0.0,
|
| 830 |
+
"batch_soft_sft_frac": 0.0,
|
| 831 |
+
"batch_tf_frac": 0.0,
|
| 832 |
+
"ce_loss": 0.0,
|
| 833 |
+
"epoch": 0.6826666666666666,
|
| 834 |
+
"grad_norm": 0.287109375,
|
| 835 |
+
"kd_loss": 0.39651421944799947,
|
| 836 |
+
"learning_rate": 3e-06,
|
| 837 |
+
"loss": 0.4956,
|
| 838 |
+
"masked_tokens": 119.6,
|
| 839 |
+
"mean_t": 0.5293499688967132,
|
| 840 |
+
"step": 320,
|
| 841 |
+
"student_masked_tokens": 119.6
|
| 842 |
+
},
|
| 843 |
+
{
|
| 844 |
+
"avg_mask_ratio": 0.5112370474264025,
|
| 845 |
+
"avg_response_length": 236.0625,
|
| 846 |
+
"avg_student_mask_ratio": 0.5112370474264025,
|
| 847 |
+
"batch_ainp_frac": 0.0,
|
| 848 |
+
"batch_inp_frac": 1.0,
|
| 849 |
+
"batch_inp_oh_frac": 0.0,
|
| 850 |
+
"batch_inp_par_frac": 0.0,
|
| 851 |
+
"batch_inp_par_par_frac": 0.0,
|
| 852 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 853 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 854 |
+
"batch_rl_frac": 0.0,
|
| 855 |
+
"batch_sft_frac": 0.0,
|
| 856 |
+
"batch_soft_sft_frac": 0.0,
|
| 857 |
+
"batch_tf_frac": 0.0,
|
| 858 |
+
"ce_loss": 0.0,
|
| 859 |
+
"epoch": 0.704,
|
| 860 |
+
"grad_norm": 0.1982421875,
|
| 861 |
+
"kd_loss": 0.4267785593967403,
|
| 862 |
+
"learning_rate": 3e-06,
|
| 863 |
+
"loss": 0.4821,
|
| 864 |
+
"masked_tokens": 129.425,
|
| 865 |
+
"mean_t": 0.5426030711154454,
|
| 866 |
+
"step": 330,
|
| 867 |
+
"student_masked_tokens": 129.425
|
| 868 |
+
},
|
| 869 |
+
{
|
| 870 |
+
"avg_mask_ratio": 0.44370225080056114,
|
| 871 |
+
"avg_response_length": 241.4875,
|
| 872 |
+
"avg_student_mask_ratio": 0.44370225080056114,
|
| 873 |
+
"batch_ainp_frac": 0.0,
|
| 874 |
+
"batch_inp_frac": 1.0,
|
| 875 |
+
"batch_inp_oh_frac": 0.0,
|
| 876 |
+
"batch_inp_par_frac": 0.0,
|
| 877 |
+
"batch_inp_par_par_frac": 0.0,
|
| 878 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 879 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 880 |
+
"batch_rl_frac": 0.0,
|
| 881 |
+
"batch_sft_frac": 0.0,
|
| 882 |
+
"batch_soft_sft_frac": 0.0,
|
| 883 |
+
"batch_tf_frac": 0.0,
|
| 884 |
+
"ce_loss": 0.0,
|
| 885 |
+
"epoch": 0.7253333333333334,
|
| 886 |
+
"grad_norm": 0.306640625,
|
| 887 |
+
"kd_loss": 0.45950967887674776,
|
| 888 |
+
"learning_rate": 3e-06,
|
| 889 |
+
"loss": 0.4426,
|
| 890 |
+
"masked_tokens": 108.775,
|
| 891 |
+
"mean_t": 0.47635243807453664,
|
| 892 |
+
"step": 340,
|
| 893 |
+
"student_masked_tokens": 108.775
|
| 894 |
+
},
|
| 895 |
+
{
|
| 896 |
+
"avg_mask_ratio": 0.49959173843380994,
|
| 897 |
+
"avg_response_length": 235.6375,
|
| 898 |
+
"avg_student_mask_ratio": 0.49959173843380994,
|
| 899 |
+
"batch_ainp_frac": 0.0,
|
| 900 |
+
"batch_inp_frac": 1.0,
|
| 901 |
+
"batch_inp_oh_frac": 0.0,
|
| 902 |
+
"batch_inp_par_frac": 0.0,
|
| 903 |
+
"batch_inp_par_par_frac": 0.0,
|
| 904 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 905 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 906 |
+
"batch_rl_frac": 0.0,
|
| 907 |
+
"batch_sft_frac": 0.0,
|
| 908 |
+
"batch_soft_sft_frac": 0.0,
|
| 909 |
+
"batch_tf_frac": 0.0,
|
| 910 |
+
"ce_loss": 0.0,
|
| 911 |
+
"epoch": 0.7466666666666667,
|
| 912 |
+
"grad_norm": 0.46875,
|
| 913 |
+
"kd_loss": 0.5008536678237376,
|
| 914 |
+
"learning_rate": 3e-06,
|
| 915 |
+
"loss": 0.5298,
|
| 916 |
+
"masked_tokens": 125.625,
|
| 917 |
+
"mean_t": 0.5268881446914747,
|
| 918 |
+
"step": 350,
|
| 919 |
+
"student_masked_tokens": 125.625
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"avg_mask_ratio": 0.4744729608530179,
|
| 923 |
+
"avg_response_length": 246.1625,
|
| 924 |
+
"avg_student_mask_ratio": 0.4744729608530179,
|
| 925 |
+
"batch_ainp_frac": 0.0,
|
| 926 |
+
"batch_inp_frac": 1.0,
|
| 927 |
+
"batch_inp_oh_frac": 0.0,
|
| 928 |
+
"batch_inp_par_frac": 0.0,
|
| 929 |
+
"batch_inp_par_par_frac": 0.0,
|
| 930 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 931 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 932 |
+
"batch_rl_frac": 0.0,
|
| 933 |
+
"batch_sft_frac": 0.0,
|
| 934 |
+
"batch_soft_sft_frac": 0.0,
|
| 935 |
+
"batch_tf_frac": 0.0,
|
| 936 |
+
"ce_loss": 0.0,
|
| 937 |
+
"epoch": 0.768,
|
| 938 |
+
"grad_norm": 0.2099609375,
|
| 939 |
+
"kd_loss": 0.5271339516211748,
|
| 940 |
+
"learning_rate": 3e-06,
|
| 941 |
+
"loss": 0.4601,
|
| 942 |
+
"masked_tokens": 116.875,
|
| 943 |
+
"mean_t": 0.5040419134311378,
|
| 944 |
+
"step": 360,
|
| 945 |
+
"student_masked_tokens": 116.875
|
| 946 |
+
},
|
| 947 |
+
{
|
| 948 |
+
"avg_mask_ratio": 0.4738045462174341,
|
| 949 |
+
"avg_response_length": 257.575,
|
| 950 |
+
"avg_student_mask_ratio": 0.4738045462174341,
|
| 951 |
+
"batch_ainp_frac": 0.0,
|
| 952 |
+
"batch_inp_frac": 1.0,
|
| 953 |
+
"batch_inp_oh_frac": 0.0,
|
| 954 |
+
"batch_inp_par_frac": 0.0,
|
| 955 |
+
"batch_inp_par_par_frac": 0.0,
|
| 956 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 957 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 958 |
+
"batch_rl_frac": 0.0,
|
| 959 |
+
"batch_sft_frac": 0.0,
|
| 960 |
+
"batch_soft_sft_frac": 0.0,
|
| 961 |
+
"batch_tf_frac": 0.0,
|
| 962 |
+
"ce_loss": 0.0,
|
| 963 |
+
"epoch": 0.7893333333333333,
|
| 964 |
+
"grad_norm": 0.166015625,
|
| 965 |
+
"kd_loss": 0.6038310030206958,
|
| 966 |
+
"learning_rate": 3e-06,
|
| 967 |
+
"loss": 0.4979,
|
| 968 |
+
"masked_tokens": 127.4625,
|
| 969 |
+
"mean_t": 0.5127181728370488,
|
| 970 |
+
"step": 370,
|
| 971 |
+
"student_masked_tokens": 127.4625
|
| 972 |
+
},
|
| 973 |
+
{
|
| 974 |
+
"avg_mask_ratio": 0.4512475330149755,
|
| 975 |
+
"avg_response_length": 209.8,
|
| 976 |
+
"avg_student_mask_ratio": 0.4512475330149755,
|
| 977 |
+
"batch_ainp_frac": 0.0,
|
| 978 |
+
"batch_inp_frac": 1.0,
|
| 979 |
+
"batch_inp_oh_frac": 0.0,
|
| 980 |
+
"batch_inp_par_frac": 0.0,
|
| 981 |
+
"batch_inp_par_par_frac": 0.0,
|
| 982 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 983 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 984 |
+
"batch_rl_frac": 0.0,
|
| 985 |
+
"batch_sft_frac": 0.0,
|
| 986 |
+
"batch_soft_sft_frac": 0.0,
|
| 987 |
+
"batch_tf_frac": 0.0,
|
| 988 |
+
"ce_loss": 0.0,
|
| 989 |
+
"epoch": 0.8106666666666666,
|
| 990 |
+
"grad_norm": 0.20703125,
|
| 991 |
+
"kd_loss": 0.4014330595062404,
|
| 992 |
+
"learning_rate": 3e-06,
|
| 993 |
+
"loss": 0.3648,
|
| 994 |
+
"masked_tokens": 100.8375,
|
| 995 |
+
"mean_t": 0.4825185665744357,
|
| 996 |
+
"step": 380,
|
| 997 |
+
"student_masked_tokens": 100.8375
|
| 998 |
+
},
|
| 999 |
+
{
|
| 1000 |
+
"avg_mask_ratio": 0.4752940105390735,
|
| 1001 |
+
"avg_response_length": 219.5625,
|
| 1002 |
+
"avg_student_mask_ratio": 0.4752940105390735,
|
| 1003 |
+
"batch_ainp_frac": 0.0,
|
| 1004 |
+
"batch_inp_frac": 1.0,
|
| 1005 |
+
"batch_inp_oh_frac": 0.0,
|
| 1006 |
+
"batch_inp_par_frac": 0.0,
|
| 1007 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1008 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1009 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1010 |
+
"batch_rl_frac": 0.0,
|
| 1011 |
+
"batch_sft_frac": 0.0,
|
| 1012 |
+
"batch_soft_sft_frac": 0.0,
|
| 1013 |
+
"batch_tf_frac": 0.0,
|
| 1014 |
+
"ce_loss": 0.0,
|
| 1015 |
+
"epoch": 0.832,
|
| 1016 |
+
"grad_norm": 0.1484375,
|
| 1017 |
+
"kd_loss": 0.4614308591259089,
|
| 1018 |
+
"learning_rate": 3e-06,
|
| 1019 |
+
"loss": 0.465,
|
| 1020 |
+
"masked_tokens": 112.375,
|
| 1021 |
+
"mean_t": 0.5053101469413377,
|
| 1022 |
+
"step": 390,
|
| 1023 |
+
"student_masked_tokens": 112.375
|
| 1024 |
+
},
|
| 1025 |
+
{
|
| 1026 |
+
"avg_mask_ratio": 0.47461870914557946,
|
| 1027 |
+
"avg_response_length": 242.6125,
|
| 1028 |
+
"avg_student_mask_ratio": 0.47461870914557946,
|
| 1029 |
+
"batch_ainp_frac": 0.0,
|
| 1030 |
+
"batch_inp_frac": 1.0,
|
| 1031 |
+
"batch_inp_oh_frac": 0.0,
|
| 1032 |
+
"batch_inp_par_frac": 0.0,
|
| 1033 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1034 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1035 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1036 |
+
"batch_rl_frac": 0.0,
|
| 1037 |
+
"batch_sft_frac": 0.0,
|
| 1038 |
+
"batch_soft_sft_frac": 0.0,
|
| 1039 |
+
"batch_tf_frac": 0.0,
|
| 1040 |
+
"ce_loss": 0.0,
|
| 1041 |
+
"epoch": 0.8533333333333334,
|
| 1042 |
+
"grad_norm": 0.3671875,
|
| 1043 |
+
"kd_loss": 0.5285846941962745,
|
| 1044 |
+
"learning_rate": 3e-06,
|
| 1045 |
+
"loss": 0.4838,
|
| 1046 |
+
"masked_tokens": 120.4125,
|
| 1047 |
+
"mean_t": 0.5052250675857067,
|
| 1048 |
+
"step": 400,
|
| 1049 |
+
"student_masked_tokens": 120.4125
|
| 1050 |
+
},
|
| 1051 |
+
{
|
| 1052 |
+
"avg_mask_ratio": 0.48321815438685006,
|
| 1053 |
+
"avg_response_length": 228.15,
|
| 1054 |
+
"avg_student_mask_ratio": 0.48321815438685006,
|
| 1055 |
+
"batch_ainp_frac": 0.0,
|
| 1056 |
+
"batch_inp_frac": 1.0,
|
| 1057 |
+
"batch_inp_oh_frac": 0.0,
|
| 1058 |
+
"batch_inp_par_frac": 0.0,
|
| 1059 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1060 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1061 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1062 |
+
"batch_rl_frac": 0.0,
|
| 1063 |
+
"batch_sft_frac": 0.0,
|
| 1064 |
+
"batch_soft_sft_frac": 0.0,
|
| 1065 |
+
"batch_tf_frac": 0.0,
|
| 1066 |
+
"ce_loss": 0.0,
|
| 1067 |
+
"epoch": 0.8746666666666667,
|
| 1068 |
+
"grad_norm": 0.2353515625,
|
| 1069 |
+
"kd_loss": 0.5044134976945088,
|
| 1070 |
+
"learning_rate": 3e-06,
|
| 1071 |
+
"loss": 0.5269,
|
| 1072 |
+
"masked_tokens": 119.0,
|
| 1073 |
+
"mean_t": 0.5050956419203431,
|
| 1074 |
+
"step": 410,
|
| 1075 |
+
"student_masked_tokens": 119.0
|
| 1076 |
+
},
|
| 1077 |
+
{
|
| 1078 |
+
"avg_mask_ratio": 0.4379329536575824,
|
| 1079 |
+
"avg_response_length": 220.225,
|
| 1080 |
+
"avg_student_mask_ratio": 0.4379329536575824,
|
| 1081 |
+
"batch_ainp_frac": 0.0,
|
| 1082 |
+
"batch_inp_frac": 1.0,
|
| 1083 |
+
"batch_inp_oh_frac": 0.0,
|
| 1084 |
+
"batch_inp_par_frac": 0.0,
|
| 1085 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1086 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1087 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1088 |
+
"batch_rl_frac": 0.0,
|
| 1089 |
+
"batch_sft_frac": 0.0,
|
| 1090 |
+
"batch_soft_sft_frac": 0.0,
|
| 1091 |
+
"batch_tf_frac": 0.0,
|
| 1092 |
+
"ce_loss": 0.0,
|
| 1093 |
+
"epoch": 0.896,
|
| 1094 |
+
"grad_norm": 0.1640625,
|
| 1095 |
+
"kd_loss": 0.27499000265762563,
|
| 1096 |
+
"learning_rate": 3e-06,
|
| 1097 |
+
"loss": 0.2998,
|
| 1098 |
+
"masked_tokens": 85.525,
|
| 1099 |
+
"mean_t": 0.4769687672611326,
|
| 1100 |
+
"step": 420,
|
| 1101 |
+
"student_masked_tokens": 85.525
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"avg_mask_ratio": 0.4674084897618741,
|
| 1105 |
+
"avg_response_length": 249.2125,
|
| 1106 |
+
"avg_student_mask_ratio": 0.4674084897618741,
|
| 1107 |
+
"batch_ainp_frac": 0.0,
|
| 1108 |
+
"batch_inp_frac": 1.0,
|
| 1109 |
+
"batch_inp_oh_frac": 0.0,
|
| 1110 |
+
"batch_inp_par_frac": 0.0,
|
| 1111 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1112 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1113 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1114 |
+
"batch_rl_frac": 0.0,
|
| 1115 |
+
"batch_sft_frac": 0.0,
|
| 1116 |
+
"batch_soft_sft_frac": 0.0,
|
| 1117 |
+
"batch_tf_frac": 0.0,
|
| 1118 |
+
"ce_loss": 0.0,
|
| 1119 |
+
"epoch": 0.9173333333333333,
|
| 1120 |
+
"grad_norm": 0.2255859375,
|
| 1121 |
+
"kd_loss": 0.4936492626361996,
|
| 1122 |
+
"learning_rate": 3e-06,
|
| 1123 |
+
"loss": 0.4862,
|
| 1124 |
+
"masked_tokens": 104.5625,
|
| 1125 |
+
"mean_t": 0.49262027950026094,
|
| 1126 |
+
"step": 430,
|
| 1127 |
+
"student_masked_tokens": 104.5625
|
| 1128 |
+
},
|
| 1129 |
+
{
|
| 1130 |
+
"avg_mask_ratio": 0.4415457699564286,
|
| 1131 |
+
"avg_response_length": 241.0875,
|
| 1132 |
+
"avg_student_mask_ratio": 0.4415457699564286,
|
| 1133 |
+
"batch_ainp_frac": 0.0,
|
| 1134 |
+
"batch_inp_frac": 1.0,
|
| 1135 |
+
"batch_inp_oh_frac": 0.0,
|
| 1136 |
+
"batch_inp_par_frac": 0.0,
|
| 1137 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1138 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1139 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1140 |
+
"batch_rl_frac": 0.0,
|
| 1141 |
+
"batch_sft_frac": 0.0,
|
| 1142 |
+
"batch_soft_sft_frac": 0.0,
|
| 1143 |
+
"batch_tf_frac": 0.0,
|
| 1144 |
+
"ce_loss": 0.0,
|
| 1145 |
+
"epoch": 0.9386666666666666,
|
| 1146 |
+
"grad_norm": 0.2265625,
|
| 1147 |
+
"kd_loss": 0.4476076794173423,
|
| 1148 |
+
"learning_rate": 3e-06,
|
| 1149 |
+
"loss": 0.4732,
|
| 1150 |
+
"masked_tokens": 113.0875,
|
| 1151 |
+
"mean_t": 0.47046207524836064,
|
| 1152 |
+
"step": 440,
|
| 1153 |
+
"student_masked_tokens": 113.0875
|
| 1154 |
+
},
|
| 1155 |
+
{
|
| 1156 |
+
"avg_mask_ratio": 0.42486972180195154,
|
| 1157 |
+
"avg_response_length": 231.9875,
|
| 1158 |
+
"avg_student_mask_ratio": 0.42486972180195154,
|
| 1159 |
+
"batch_ainp_frac": 0.0,
|
| 1160 |
+
"batch_inp_frac": 1.0,
|
| 1161 |
+
"batch_inp_oh_frac": 0.0,
|
| 1162 |
+
"batch_inp_par_frac": 0.0,
|
| 1163 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1164 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1165 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1166 |
+
"batch_rl_frac": 0.0,
|
| 1167 |
+
"batch_sft_frac": 0.0,
|
| 1168 |
+
"batch_soft_sft_frac": 0.0,
|
| 1169 |
+
"batch_tf_frac": 0.0,
|
| 1170 |
+
"ce_loss": 0.0,
|
| 1171 |
+
"epoch": 0.96,
|
| 1172 |
+
"grad_norm": 0.255859375,
|
| 1173 |
+
"kd_loss": 0.40184078344118745,
|
| 1174 |
+
"learning_rate": 3e-06,
|
| 1175 |
+
"loss": 0.3814,
|
| 1176 |
+
"masked_tokens": 103.4,
|
| 1177 |
+
"mean_t": 0.45781184462830427,
|
| 1178 |
+
"step": 450,
|
| 1179 |
+
"student_masked_tokens": 103.4
|
| 1180 |
+
},
|
| 1181 |
+
{
|
| 1182 |
+
"avg_mask_ratio": 0.47578654896933587,
|
| 1183 |
+
"avg_response_length": 214.6125,
|
| 1184 |
+
"avg_student_mask_ratio": 0.47578654896933587,
|
| 1185 |
+
"batch_ainp_frac": 0.0,
|
| 1186 |
+
"batch_inp_frac": 1.0,
|
| 1187 |
+
"batch_inp_oh_frac": 0.0,
|
| 1188 |
+
"batch_inp_par_frac": 0.0,
|
| 1189 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1190 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1191 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1192 |
+
"batch_rl_frac": 0.0,
|
| 1193 |
+
"batch_sft_frac": 0.0,
|
| 1194 |
+
"batch_soft_sft_frac": 0.0,
|
| 1195 |
+
"batch_tf_frac": 0.0,
|
| 1196 |
+
"ce_loss": 0.0,
|
| 1197 |
+
"epoch": 0.9813333333333333,
|
| 1198 |
+
"grad_norm": 0.15625,
|
| 1199 |
+
"kd_loss": 0.43999221579811093,
|
| 1200 |
+
"learning_rate": 3e-06,
|
| 1201 |
+
"loss": 0.45,
|
| 1202 |
+
"masked_tokens": 105.3125,
|
| 1203 |
+
"mean_t": 0.5075790266972036,
|
| 1204 |
+
"step": 460,
|
| 1205 |
+
"student_masked_tokens": 105.3125
|
| 1206 |
+
},
|
| 1207 |
+
{
|
| 1208 |
+
"avg_mask_ratio": 0.4782901787132557,
|
| 1209 |
+
"avg_response_length": 224.0952380952381,
|
| 1210 |
+
"avg_student_mask_ratio": 0.4782901787132557,
|
| 1211 |
+
"batch_ainp_frac": 0.0,
|
| 1212 |
+
"batch_inp_frac": 1.0,
|
| 1213 |
+
"batch_inp_oh_frac": 0.0,
|
| 1214 |
+
"batch_inp_par_frac": 0.0,
|
| 1215 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1216 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1217 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1218 |
+
"batch_rl_frac": 0.0,
|
| 1219 |
+
"batch_sft_frac": 0.0,
|
| 1220 |
+
"batch_soft_sft_frac": 0.0,
|
| 1221 |
+
"batch_tf_frac": 0.0,
|
| 1222 |
+
"ce_loss": 0.0,
|
| 1223 |
+
"epoch": 1.0042666666666666,
|
| 1224 |
+
"grad_norm": 0.330078125,
|
| 1225 |
+
"kd_loss": 0.5121579658046268,
|
| 1226 |
+
"learning_rate": 3e-06,
|
| 1227 |
+
"loss": 0.5319,
|
| 1228 |
+
"masked_tokens": 107.23809523809524,
|
| 1229 |
+
"mean_t": 0.5031429776822084,
|
| 1230 |
+
"step": 470,
|
| 1231 |
+
"student_masked_tokens": 107.23809523809524
|
| 1232 |
+
},
|
| 1233 |
+
{
|
| 1234 |
+
"avg_mask_ratio": 0.47575968883465974,
|
| 1235 |
+
"avg_response_length": 249.4125,
|
| 1236 |
+
"avg_student_mask_ratio": 0.47575968883465974,
|
| 1237 |
+
"batch_ainp_frac": 0.0,
|
| 1238 |
+
"batch_inp_frac": 1.0,
|
| 1239 |
+
"batch_inp_oh_frac": 0.0,
|
| 1240 |
+
"batch_inp_par_frac": 0.0,
|
| 1241 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1242 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1243 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1244 |
+
"batch_rl_frac": 0.0,
|
| 1245 |
+
"batch_sft_frac": 0.0,
|
| 1246 |
+
"batch_soft_sft_frac": 0.0,
|
| 1247 |
+
"batch_tf_frac": 0.0,
|
| 1248 |
+
"ce_loss": 0.0,
|
| 1249 |
+
"epoch": 1.0256,
|
| 1250 |
+
"grad_norm": 0.220703125,
|
| 1251 |
+
"kd_loss": 0.5364690735516149,
|
| 1252 |
+
"learning_rate": 3e-06,
|
| 1253 |
+
"loss": 0.4899,
|
| 1254 |
+
"masked_tokens": 118.35,
|
| 1255 |
+
"mean_t": 0.504472183593316,
|
| 1256 |
+
"step": 480,
|
| 1257 |
+
"student_masked_tokens": 118.35
|
| 1258 |
+
},
|
| 1259 |
+
{
|
| 1260 |
+
"avg_mask_ratio": 0.4563717324635945,
|
| 1261 |
+
"avg_response_length": 232.0375,
|
| 1262 |
+
"avg_student_mask_ratio": 0.4563717324635945,
|
| 1263 |
+
"batch_ainp_frac": 0.0,
|
| 1264 |
+
"batch_inp_frac": 1.0,
|
| 1265 |
+
"batch_inp_oh_frac": 0.0,
|
| 1266 |
+
"batch_inp_par_frac": 0.0,
|
| 1267 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1268 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1269 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1270 |
+
"batch_rl_frac": 0.0,
|
| 1271 |
+
"batch_sft_frac": 0.0,
|
| 1272 |
+
"batch_soft_sft_frac": 0.0,
|
| 1273 |
+
"batch_tf_frac": 0.0,
|
| 1274 |
+
"ce_loss": 0.0,
|
| 1275 |
+
"epoch": 1.0469333333333333,
|
| 1276 |
+
"grad_norm": 0.39453125,
|
| 1277 |
+
"kd_loss": 0.38637587467336515,
|
| 1278 |
+
"learning_rate": 3e-06,
|
| 1279 |
+
"loss": 0.4045,
|
| 1280 |
+
"masked_tokens": 98.35,
|
| 1281 |
+
"mean_t": 0.4888980514719151,
|
| 1282 |
+
"step": 490,
|
| 1283 |
+
"student_masked_tokens": 98.35
|
| 1284 |
+
},
|
| 1285 |
+
{
|
| 1286 |
+
"avg_mask_ratio": 0.5079968665260821,
|
| 1287 |
+
"avg_response_length": 253.7875,
|
| 1288 |
+
"avg_student_mask_ratio": 0.5079968665260821,
|
| 1289 |
+
"batch_ainp_frac": 0.0,
|
| 1290 |
+
"batch_inp_frac": 1.0,
|
| 1291 |
+
"batch_inp_oh_frac": 0.0,
|
| 1292 |
+
"batch_inp_par_frac": 0.0,
|
| 1293 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1294 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1295 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1296 |
+
"batch_rl_frac": 0.0,
|
| 1297 |
+
"batch_sft_frac": 0.0,
|
| 1298 |
+
"batch_soft_sft_frac": 0.0,
|
| 1299 |
+
"batch_tf_frac": 0.0,
|
| 1300 |
+
"ce_loss": 0.0,
|
| 1301 |
+
"epoch": 1.0682666666666667,
|
| 1302 |
+
"grad_norm": 0.26171875,
|
| 1303 |
+
"kd_loss": 0.4493013881629054,
|
| 1304 |
+
"learning_rate": 3e-06,
|
| 1305 |
+
"loss": 0.4348,
|
| 1306 |
+
"masked_tokens": 128.225,
|
| 1307 |
+
"mean_t": 0.5469163245841628,
|
| 1308 |
+
"step": 500,
|
| 1309 |
+
"student_masked_tokens": 128.225
|
| 1310 |
+
},
|
| 1311 |
+
{
|
| 1312 |
+
"avg_mask_ratio": 0.5109448074479588,
|
| 1313 |
+
"avg_response_length": 254.2,
|
| 1314 |
+
"avg_student_mask_ratio": 0.5109448074479588,
|
| 1315 |
+
"batch_ainp_frac": 0.0,
|
| 1316 |
+
"batch_inp_frac": 1.0,
|
| 1317 |
+
"batch_inp_oh_frac": 0.0,
|
| 1318 |
+
"batch_inp_par_frac": 0.0,
|
| 1319 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1320 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1321 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1322 |
+
"batch_rl_frac": 0.0,
|
| 1323 |
+
"batch_sft_frac": 0.0,
|
| 1324 |
+
"batch_soft_sft_frac": 0.0,
|
| 1325 |
+
"batch_tf_frac": 0.0,
|
| 1326 |
+
"ce_loss": 0.0,
|
| 1327 |
+
"epoch": 1.0896,
|
| 1328 |
+
"grad_norm": 0.2255859375,
|
| 1329 |
+
"kd_loss": 0.5702315452758938,
|
| 1330 |
+
"learning_rate": 3e-06,
|
| 1331 |
+
"loss": 0.5036,
|
| 1332 |
+
"masked_tokens": 137.5875,
|
| 1333 |
+
"mean_t": 0.5275314710394013,
|
| 1334 |
+
"step": 510,
|
| 1335 |
+
"student_masked_tokens": 137.5875
|
| 1336 |
+
},
|
| 1337 |
+
{
|
| 1338 |
+
"avg_mask_ratio": 0.45396183808334173,
|
| 1339 |
+
"avg_response_length": 202.7625,
|
| 1340 |
+
"avg_student_mask_ratio": 0.45396183808334173,
|
| 1341 |
+
"batch_ainp_frac": 0.0,
|
| 1342 |
+
"batch_inp_frac": 1.0,
|
| 1343 |
+
"batch_inp_oh_frac": 0.0,
|
| 1344 |
+
"batch_inp_par_frac": 0.0,
|
| 1345 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1346 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1347 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1348 |
+
"batch_rl_frac": 0.0,
|
| 1349 |
+
"batch_sft_frac": 0.0,
|
| 1350 |
+
"batch_soft_sft_frac": 0.0,
|
| 1351 |
+
"batch_tf_frac": 0.0,
|
| 1352 |
+
"ce_loss": 0.0,
|
| 1353 |
+
"epoch": 1.1109333333333333,
|
| 1354 |
+
"grad_norm": 0.2421875,
|
| 1355 |
+
"kd_loss": 0.420669932495889,
|
| 1356 |
+
"learning_rate": 3e-06,
|
| 1357 |
+
"loss": 0.4283,
|
| 1358 |
+
"masked_tokens": 87.0625,
|
| 1359 |
+
"mean_t": 0.49193521235138177,
|
| 1360 |
+
"step": 520,
|
| 1361 |
+
"student_masked_tokens": 87.0625
|
| 1362 |
+
},
|
| 1363 |
+
{
|
| 1364 |
+
"avg_mask_ratio": 0.47015948037151245,
|
| 1365 |
+
"avg_response_length": 214.275,
|
| 1366 |
+
"avg_student_mask_ratio": 0.47015948037151245,
|
| 1367 |
+
"batch_ainp_frac": 0.0,
|
| 1368 |
+
"batch_inp_frac": 1.0,
|
| 1369 |
+
"batch_inp_oh_frac": 0.0,
|
| 1370 |
+
"batch_inp_par_frac": 0.0,
|
| 1371 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1372 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1373 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1374 |
+
"batch_rl_frac": 0.0,
|
| 1375 |
+
"batch_sft_frac": 0.0,
|
| 1376 |
+
"batch_soft_sft_frac": 0.0,
|
| 1377 |
+
"batch_tf_frac": 0.0,
|
| 1378 |
+
"ce_loss": 0.0,
|
| 1379 |
+
"epoch": 1.1322666666666668,
|
| 1380 |
+
"grad_norm": 0.31640625,
|
| 1381 |
+
"kd_loss": 0.4567407854818356,
|
| 1382 |
+
"learning_rate": 3e-06,
|
| 1383 |
+
"loss": 0.5135,
|
| 1384 |
+
"masked_tokens": 99.8625,
|
| 1385 |
+
"mean_t": 0.49621942077938,
|
| 1386 |
+
"step": 530,
|
| 1387 |
+
"student_masked_tokens": 99.8625
|
| 1388 |
+
},
|
| 1389 |
+
{
|
| 1390 |
+
"avg_mask_ratio": 0.4892866689246148,
|
| 1391 |
+
"avg_response_length": 231.3125,
|
| 1392 |
+
"avg_student_mask_ratio": 0.4892866689246148,
|
| 1393 |
+
"batch_ainp_frac": 0.0,
|
| 1394 |
+
"batch_inp_frac": 1.0,
|
| 1395 |
+
"batch_inp_oh_frac": 0.0,
|
| 1396 |
+
"batch_inp_par_frac": 0.0,
|
| 1397 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1398 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1399 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1400 |
+
"batch_rl_frac": 0.0,
|
| 1401 |
+
"batch_sft_frac": 0.0,
|
| 1402 |
+
"batch_soft_sft_frac": 0.0,
|
| 1403 |
+
"batch_tf_frac": 0.0,
|
| 1404 |
+
"ce_loss": 0.0,
|
| 1405 |
+
"epoch": 1.1536,
|
| 1406 |
+
"grad_norm": 0.3359375,
|
| 1407 |
+
"kd_loss": 0.5524838666212872,
|
| 1408 |
+
"learning_rate": 3e-06,
|
| 1409 |
+
"loss": 0.5048,
|
| 1410 |
+
"masked_tokens": 107.375,
|
| 1411 |
+
"mean_t": 0.5208023569080978,
|
| 1412 |
+
"step": 540,
|
| 1413 |
+
"student_masked_tokens": 107.375
|
| 1414 |
+
},
|
| 1415 |
+
{
|
| 1416 |
+
"avg_mask_ratio": 0.4541942774085328,
|
| 1417 |
+
"avg_response_length": 213.525,
|
| 1418 |
+
"avg_student_mask_ratio": 0.4541942774085328,
|
| 1419 |
+
"batch_ainp_frac": 0.0,
|
| 1420 |
+
"batch_inp_frac": 1.0,
|
| 1421 |
+
"batch_inp_oh_frac": 0.0,
|
| 1422 |
+
"batch_inp_par_frac": 0.0,
|
| 1423 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1424 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1425 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1426 |
+
"batch_rl_frac": 0.0,
|
| 1427 |
+
"batch_sft_frac": 0.0,
|
| 1428 |
+
"batch_soft_sft_frac": 0.0,
|
| 1429 |
+
"batch_tf_frac": 0.0,
|
| 1430 |
+
"ce_loss": 0.0,
|
| 1431 |
+
"epoch": 1.1749333333333334,
|
| 1432 |
+
"grad_norm": 0.1572265625,
|
| 1433 |
+
"kd_loss": 0.3688938313097424,
|
| 1434 |
+
"learning_rate": 3e-06,
|
| 1435 |
+
"loss": 0.3777,
|
| 1436 |
+
"masked_tokens": 97.8875,
|
| 1437 |
+
"mean_t": 0.4767197913257405,
|
| 1438 |
+
"step": 550,
|
| 1439 |
+
"student_masked_tokens": 97.8875
|
| 1440 |
+
},
|
| 1441 |
+
{
|
| 1442 |
+
"avg_mask_ratio": 0.39282396506750955,
|
| 1443 |
+
"avg_response_length": 231.4125,
|
| 1444 |
+
"avg_student_mask_ratio": 0.39282396506750955,
|
| 1445 |
+
"batch_ainp_frac": 0.0,
|
| 1446 |
+
"batch_inp_frac": 1.0,
|
| 1447 |
+
"batch_inp_oh_frac": 0.0,
|
| 1448 |
+
"batch_inp_par_frac": 0.0,
|
| 1449 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1450 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1451 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1452 |
+
"batch_rl_frac": 0.0,
|
| 1453 |
+
"batch_sft_frac": 0.0,
|
| 1454 |
+
"batch_soft_sft_frac": 0.0,
|
| 1455 |
+
"batch_tf_frac": 0.0,
|
| 1456 |
+
"ce_loss": 0.0,
|
| 1457 |
+
"epoch": 1.1962666666666666,
|
| 1458 |
+
"grad_norm": 0.359375,
|
| 1459 |
+
"kd_loss": 0.4847641052244853,
|
| 1460 |
+
"learning_rate": 3e-06,
|
| 1461 |
+
"loss": 0.4135,
|
| 1462 |
+
"masked_tokens": 99.2875,
|
| 1463 |
+
"mean_t": 0.4111072298779618,
|
| 1464 |
+
"step": 560,
|
| 1465 |
+
"student_masked_tokens": 99.2875
|
| 1466 |
+
},
|
| 1467 |
+
{
|
| 1468 |
+
"avg_mask_ratio": 0.4483634108910337,
|
| 1469 |
+
"avg_response_length": 230.1625,
|
| 1470 |
+
"avg_student_mask_ratio": 0.4483634108910337,
|
| 1471 |
+
"batch_ainp_frac": 0.0,
|
| 1472 |
+
"batch_inp_frac": 1.0,
|
| 1473 |
+
"batch_inp_oh_frac": 0.0,
|
| 1474 |
+
"batch_inp_par_frac": 0.0,
|
| 1475 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1476 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1477 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1478 |
+
"batch_rl_frac": 0.0,
|
| 1479 |
+
"batch_sft_frac": 0.0,
|
| 1480 |
+
"batch_soft_sft_frac": 0.0,
|
| 1481 |
+
"batch_tf_frac": 0.0,
|
| 1482 |
+
"ce_loss": 0.0,
|
| 1483 |
+
"epoch": 1.2176,
|
| 1484 |
+
"grad_norm": 0.2294921875,
|
| 1485 |
+
"kd_loss": 0.41131700558344164,
|
| 1486 |
+
"learning_rate": 3e-06,
|
| 1487 |
+
"loss": 0.4323,
|
| 1488 |
+
"masked_tokens": 108.9875,
|
| 1489 |
+
"mean_t": 0.48533305872697385,
|
| 1490 |
+
"step": 570,
|
| 1491 |
+
"student_masked_tokens": 108.9875
|
| 1492 |
+
},
|
| 1493 |
+
{
|
| 1494 |
+
"avg_mask_ratio": 0.465452536707744,
|
| 1495 |
+
"avg_response_length": 267.4375,
|
| 1496 |
+
"avg_student_mask_ratio": 0.465452536707744,
|
| 1497 |
+
"batch_ainp_frac": 0.0,
|
| 1498 |
+
"batch_inp_frac": 1.0,
|
| 1499 |
+
"batch_inp_oh_frac": 0.0,
|
| 1500 |
+
"batch_inp_par_frac": 0.0,
|
| 1501 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1502 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1503 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1504 |
+
"batch_rl_frac": 0.0,
|
| 1505 |
+
"batch_sft_frac": 0.0,
|
| 1506 |
+
"batch_soft_sft_frac": 0.0,
|
| 1507 |
+
"batch_tf_frac": 0.0,
|
| 1508 |
+
"ce_loss": 0.0,
|
| 1509 |
+
"epoch": 1.2389333333333332,
|
| 1510 |
+
"grad_norm": 0.251953125,
|
| 1511 |
+
"kd_loss": 0.44281044333342834,
|
| 1512 |
+
"learning_rate": 3e-06,
|
| 1513 |
+
"loss": 0.4225,
|
| 1514 |
+
"masked_tokens": 129.075,
|
| 1515 |
+
"mean_t": 0.49976949762785805,
|
| 1516 |
+
"step": 580,
|
| 1517 |
+
"student_masked_tokens": 129.075
|
| 1518 |
+
},
|
| 1519 |
+
{
|
| 1520 |
+
"avg_mask_ratio": 0.5225977989146486,
|
| 1521 |
+
"avg_response_length": 228.45,
|
| 1522 |
+
"avg_student_mask_ratio": 0.5225977989146486,
|
| 1523 |
+
"batch_ainp_frac": 0.0,
|
| 1524 |
+
"batch_inp_frac": 1.0,
|
| 1525 |
+
"batch_inp_oh_frac": 0.0,
|
| 1526 |
+
"batch_inp_par_frac": 0.0,
|
| 1527 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1528 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1529 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1530 |
+
"batch_rl_frac": 0.0,
|
| 1531 |
+
"batch_sft_frac": 0.0,
|
| 1532 |
+
"batch_soft_sft_frac": 0.0,
|
| 1533 |
+
"batch_tf_frac": 0.0,
|
| 1534 |
+
"ce_loss": 0.0,
|
| 1535 |
+
"epoch": 1.2602666666666666,
|
| 1536 |
+
"grad_norm": 0.251953125,
|
| 1537 |
+
"kd_loss": 0.5374341921963897,
|
| 1538 |
+
"learning_rate": 3e-06,
|
| 1539 |
+
"loss": 0.5561,
|
| 1540 |
+
"masked_tokens": 121.675,
|
| 1541 |
+
"mean_t": 0.5521843038732186,
|
| 1542 |
+
"step": 590,
|
| 1543 |
+
"student_masked_tokens": 121.675
|
| 1544 |
+
},
|
| 1545 |
+
{
|
| 1546 |
+
"avg_mask_ratio": 0.44998724836623294,
|
| 1547 |
+
"avg_response_length": 236.7,
|
| 1548 |
+
"avg_student_mask_ratio": 0.44998724836623294,
|
| 1549 |
+
"batch_ainp_frac": 0.0,
|
| 1550 |
+
"batch_inp_frac": 1.0,
|
| 1551 |
+
"batch_inp_oh_frac": 0.0,
|
| 1552 |
+
"batch_inp_par_frac": 0.0,
|
| 1553 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1554 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1555 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1556 |
+
"batch_rl_frac": 0.0,
|
| 1557 |
+
"batch_sft_frac": 0.0,
|
| 1558 |
+
"batch_soft_sft_frac": 0.0,
|
| 1559 |
+
"batch_tf_frac": 0.0,
|
| 1560 |
+
"ce_loss": 0.0,
|
| 1561 |
+
"epoch": 1.2816,
|
| 1562 |
+
"grad_norm": 0.2392578125,
|
| 1563 |
+
"kd_loss": 0.41848757546744225,
|
| 1564 |
+
"learning_rate": 3e-06,
|
| 1565 |
+
"loss": 0.4048,
|
| 1566 |
+
"masked_tokens": 110.1625,
|
| 1567 |
+
"mean_t": 0.4788527532829903,
|
| 1568 |
+
"step": 600,
|
| 1569 |
+
"student_masked_tokens": 110.1625
|
| 1570 |
+
},
|
| 1571 |
+
{
|
| 1572 |
+
"avg_mask_ratio": 0.5042130865273066,
|
| 1573 |
+
"avg_response_length": 230.3375,
|
| 1574 |
+
"avg_student_mask_ratio": 0.5042130865273066,
|
| 1575 |
+
"batch_ainp_frac": 0.0,
|
| 1576 |
+
"batch_inp_frac": 1.0,
|
| 1577 |
+
"batch_inp_oh_frac": 0.0,
|
| 1578 |
+
"batch_inp_par_frac": 0.0,
|
| 1579 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1580 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1581 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1582 |
+
"batch_rl_frac": 0.0,
|
| 1583 |
+
"batch_sft_frac": 0.0,
|
| 1584 |
+
"batch_soft_sft_frac": 0.0,
|
| 1585 |
+
"batch_tf_frac": 0.0,
|
| 1586 |
+
"ce_loss": 0.0,
|
| 1587 |
+
"epoch": 1.3029333333333333,
|
| 1588 |
+
"grad_norm": 0.255859375,
|
| 1589 |
+
"kd_loss": 0.5498438457787416,
|
| 1590 |
+
"learning_rate": 3e-06,
|
| 1591 |
+
"loss": 0.5621,
|
| 1592 |
+
"masked_tokens": 112.625,
|
| 1593 |
+
"mean_t": 0.5445419924799353,
|
| 1594 |
+
"step": 610,
|
| 1595 |
+
"student_masked_tokens": 112.625
|
| 1596 |
+
},
|
| 1597 |
+
{
|
| 1598 |
+
"avg_mask_ratio": 0.49637898594373836,
|
| 1599 |
+
"avg_response_length": 233.0625,
|
| 1600 |
+
"avg_student_mask_ratio": 0.49637898594373836,
|
| 1601 |
+
"batch_ainp_frac": 0.0,
|
| 1602 |
+
"batch_inp_frac": 1.0,
|
| 1603 |
+
"batch_inp_oh_frac": 0.0,
|
| 1604 |
+
"batch_inp_par_frac": 0.0,
|
| 1605 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1606 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1607 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1608 |
+
"batch_rl_frac": 0.0,
|
| 1609 |
+
"batch_sft_frac": 0.0,
|
| 1610 |
+
"batch_soft_sft_frac": 0.0,
|
| 1611 |
+
"batch_tf_frac": 0.0,
|
| 1612 |
+
"ce_loss": 0.0,
|
| 1613 |
+
"epoch": 1.3242666666666667,
|
| 1614 |
+
"grad_norm": 0.2060546875,
|
| 1615 |
+
"kd_loss": 0.5435912919167265,
|
| 1616 |
+
"learning_rate": 3e-06,
|
| 1617 |
+
"loss": 0.5191,
|
| 1618 |
+
"masked_tokens": 111.25,
|
| 1619 |
+
"mean_t": 0.52490478400141,
|
| 1620 |
+
"step": 620,
|
| 1621 |
+
"student_masked_tokens": 111.25
|
| 1622 |
+
},
|
| 1623 |
+
{
|
| 1624 |
+
"avg_mask_ratio": 0.5177568581304512,
|
| 1625 |
+
"avg_response_length": 257.2125,
|
| 1626 |
+
"avg_student_mask_ratio": 0.5177568581304512,
|
| 1627 |
+
"batch_ainp_frac": 0.0,
|
| 1628 |
+
"batch_inp_frac": 1.0,
|
| 1629 |
+
"batch_inp_oh_frac": 0.0,
|
| 1630 |
+
"batch_inp_par_frac": 0.0,
|
| 1631 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1632 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1633 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1634 |
+
"batch_rl_frac": 0.0,
|
| 1635 |
+
"batch_sft_frac": 0.0,
|
| 1636 |
+
"batch_soft_sft_frac": 0.0,
|
| 1637 |
+
"batch_tf_frac": 0.0,
|
| 1638 |
+
"ce_loss": 0.0,
|
| 1639 |
+
"epoch": 1.3456000000000001,
|
| 1640 |
+
"grad_norm": 0.423828125,
|
| 1641 |
+
"kd_loss": 0.534644598815612,
|
| 1642 |
+
"learning_rate": 3e-06,
|
| 1643 |
+
"loss": 0.5523,
|
| 1644 |
+
"masked_tokens": 129.6125,
|
| 1645 |
+
"mean_t": 0.5564947265549562,
|
| 1646 |
+
"step": 630,
|
| 1647 |
+
"student_masked_tokens": 129.6125
|
| 1648 |
+
},
|
| 1649 |
+
{
|
| 1650 |
+
"avg_mask_ratio": 0.48226998368045315,
|
| 1651 |
+
"avg_response_length": 237.7125,
|
| 1652 |
+
"avg_student_mask_ratio": 0.48226998368045315,
|
| 1653 |
+
"batch_ainp_frac": 0.0,
|
| 1654 |
+
"batch_inp_frac": 1.0,
|
| 1655 |
+
"batch_inp_oh_frac": 0.0,
|
| 1656 |
+
"batch_inp_par_frac": 0.0,
|
| 1657 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1658 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1659 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1660 |
+
"batch_rl_frac": 0.0,
|
| 1661 |
+
"batch_sft_frac": 0.0,
|
| 1662 |
+
"batch_soft_sft_frac": 0.0,
|
| 1663 |
+
"batch_tf_frac": 0.0,
|
| 1664 |
+
"ce_loss": 0.0,
|
| 1665 |
+
"epoch": 1.3669333333333333,
|
| 1666 |
+
"grad_norm": 0.138671875,
|
| 1667 |
+
"kd_loss": 0.363254164889031,
|
| 1668 |
+
"learning_rate": 3e-06,
|
| 1669 |
+
"loss": 0.3199,
|
| 1670 |
+
"masked_tokens": 120.275,
|
| 1671 |
+
"mean_t": 0.5093393943971023,
|
| 1672 |
+
"step": 640,
|
| 1673 |
+
"student_masked_tokens": 120.275
|
| 1674 |
+
},
|
| 1675 |
+
{
|
| 1676 |
+
"avg_mask_ratio": 0.5306948523037136,
|
| 1677 |
+
"avg_response_length": 238.0125,
|
| 1678 |
+
"avg_student_mask_ratio": 0.5306948523037136,
|
| 1679 |
+
"batch_ainp_frac": 0.0,
|
| 1680 |
+
"batch_inp_frac": 1.0,
|
| 1681 |
+
"batch_inp_oh_frac": 0.0,
|
| 1682 |
+
"batch_inp_par_frac": 0.0,
|
| 1683 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1684 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1685 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1686 |
+
"batch_rl_frac": 0.0,
|
| 1687 |
+
"batch_sft_frac": 0.0,
|
| 1688 |
+
"batch_soft_sft_frac": 0.0,
|
| 1689 |
+
"batch_tf_frac": 0.0,
|
| 1690 |
+
"ce_loss": 0.0,
|
| 1691 |
+
"epoch": 1.3882666666666665,
|
| 1692 |
+
"grad_norm": 0.158203125,
|
| 1693 |
+
"kd_loss": 0.5049424907911544,
|
| 1694 |
+
"learning_rate": 3e-06,
|
| 1695 |
+
"loss": 0.4932,
|
| 1696 |
+
"masked_tokens": 133.25,
|
| 1697 |
+
"mean_t": 0.5558586571365595,
|
| 1698 |
+
"step": 650,
|
| 1699 |
+
"student_masked_tokens": 133.25
|
| 1700 |
+
},
|
| 1701 |
+
{
|
| 1702 |
+
"avg_mask_ratio": 0.4821273953886703,
|
| 1703 |
+
"avg_response_length": 247.775,
|
| 1704 |
+
"avg_student_mask_ratio": 0.4821273953886703,
|
| 1705 |
+
"batch_ainp_frac": 0.0,
|
| 1706 |
+
"batch_inp_frac": 1.0,
|
| 1707 |
+
"batch_inp_oh_frac": 0.0,
|
| 1708 |
+
"batch_inp_par_frac": 0.0,
|
| 1709 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1710 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1711 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1712 |
+
"batch_rl_frac": 0.0,
|
| 1713 |
+
"batch_sft_frac": 0.0,
|
| 1714 |
+
"batch_soft_sft_frac": 0.0,
|
| 1715 |
+
"batch_tf_frac": 0.0,
|
| 1716 |
+
"ce_loss": 0.0,
|
| 1717 |
+
"epoch": 1.4096,
|
| 1718 |
+
"grad_norm": 0.23828125,
|
| 1719 |
+
"kd_loss": 0.41951912807553526,
|
| 1720 |
+
"learning_rate": 3e-06,
|
| 1721 |
+
"loss": 0.4242,
|
| 1722 |
+
"masked_tokens": 128.875,
|
| 1723 |
+
"mean_t": 0.51307404555846,
|
| 1724 |
+
"step": 660,
|
| 1725 |
+
"student_masked_tokens": 128.875
|
| 1726 |
+
},
|
| 1727 |
+
{
|
| 1728 |
+
"avg_mask_ratio": 0.46056515555246735,
|
| 1729 |
+
"avg_response_length": 240.4375,
|
| 1730 |
+
"avg_student_mask_ratio": 0.46056515555246735,
|
| 1731 |
+
"batch_ainp_frac": 0.0,
|
| 1732 |
+
"batch_inp_frac": 1.0,
|
| 1733 |
+
"batch_inp_oh_frac": 0.0,
|
| 1734 |
+
"batch_inp_par_frac": 0.0,
|
| 1735 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1736 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1737 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1738 |
+
"batch_rl_frac": 0.0,
|
| 1739 |
+
"batch_sft_frac": 0.0,
|
| 1740 |
+
"batch_soft_sft_frac": 0.0,
|
| 1741 |
+
"batch_tf_frac": 0.0,
|
| 1742 |
+
"ce_loss": 0.0,
|
| 1743 |
+
"epoch": 1.4309333333333334,
|
| 1744 |
+
"grad_norm": 0.28515625,
|
| 1745 |
+
"kd_loss": 0.3502249997708077,
|
| 1746 |
+
"learning_rate": 3e-06,
|
| 1747 |
+
"loss": 0.383,
|
| 1748 |
+
"masked_tokens": 119.85,
|
| 1749 |
+
"mean_t": 0.4907285622088239,
|
| 1750 |
+
"step": 670,
|
| 1751 |
+
"student_masked_tokens": 119.85
|
| 1752 |
+
},
|
| 1753 |
+
{
|
| 1754 |
+
"avg_mask_ratio": 0.4666106043441687,
|
| 1755 |
+
"avg_response_length": 226.7375,
|
| 1756 |
+
"avg_student_mask_ratio": 0.4666106043441687,
|
| 1757 |
+
"batch_ainp_frac": 0.0,
|
| 1758 |
+
"batch_inp_frac": 1.0,
|
| 1759 |
+
"batch_inp_oh_frac": 0.0,
|
| 1760 |
+
"batch_inp_par_frac": 0.0,
|
| 1761 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1762 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1763 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1764 |
+
"batch_rl_frac": 0.0,
|
| 1765 |
+
"batch_sft_frac": 0.0,
|
| 1766 |
+
"batch_soft_sft_frac": 0.0,
|
| 1767 |
+
"batch_tf_frac": 0.0,
|
| 1768 |
+
"ce_loss": 0.0,
|
| 1769 |
+
"epoch": 1.4522666666666666,
|
| 1770 |
+
"grad_norm": 0.3828125,
|
| 1771 |
+
"kd_loss": 0.4872104180909183,
|
| 1772 |
+
"learning_rate": 3e-06,
|
| 1773 |
+
"loss": 0.4906,
|
| 1774 |
+
"masked_tokens": 107.4625,
|
| 1775 |
+
"mean_t": 0.4913603452499956,
|
| 1776 |
+
"step": 680,
|
| 1777 |
+
"student_masked_tokens": 107.4625
|
| 1778 |
+
},
|
| 1779 |
+
{
|
| 1780 |
+
"avg_mask_ratio": 0.4790851596510038,
|
| 1781 |
+
"avg_response_length": 202.05,
|
| 1782 |
+
"avg_student_mask_ratio": 0.4790851596510038,
|
| 1783 |
+
"batch_ainp_frac": 0.0,
|
| 1784 |
+
"batch_inp_frac": 1.0,
|
| 1785 |
+
"batch_inp_oh_frac": 0.0,
|
| 1786 |
+
"batch_inp_par_frac": 0.0,
|
| 1787 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1788 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1789 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1790 |
+
"batch_rl_frac": 0.0,
|
| 1791 |
+
"batch_sft_frac": 0.0,
|
| 1792 |
+
"batch_soft_sft_frac": 0.0,
|
| 1793 |
+
"batch_tf_frac": 0.0,
|
| 1794 |
+
"ce_loss": 0.0,
|
| 1795 |
+
"epoch": 1.4736,
|
| 1796 |
+
"grad_norm": 0.2197265625,
|
| 1797 |
+
"kd_loss": 0.41381314450108847,
|
| 1798 |
+
"learning_rate": 3e-06,
|
| 1799 |
+
"loss": 0.453,
|
| 1800 |
+
"masked_tokens": 111.3125,
|
| 1801 |
+
"mean_t": 0.5133644798654131,
|
| 1802 |
+
"step": 690,
|
| 1803 |
+
"student_masked_tokens": 111.3125
|
| 1804 |
+
},
|
| 1805 |
+
{
|
| 1806 |
+
"avg_mask_ratio": 0.5250519359949977,
|
| 1807 |
+
"avg_response_length": 228.125,
|
| 1808 |
+
"avg_student_mask_ratio": 0.5250519359949977,
|
| 1809 |
+
"batch_ainp_frac": 0.0,
|
| 1810 |
+
"batch_inp_frac": 1.0,
|
| 1811 |
+
"batch_inp_oh_frac": 0.0,
|
| 1812 |
+
"batch_inp_par_frac": 0.0,
|
| 1813 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1814 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1815 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1816 |
+
"batch_rl_frac": 0.0,
|
| 1817 |
+
"batch_sft_frac": 0.0,
|
| 1818 |
+
"batch_soft_sft_frac": 0.0,
|
| 1819 |
+
"batch_tf_frac": 0.0,
|
| 1820 |
+
"ce_loss": 0.0,
|
| 1821 |
+
"epoch": 1.4949333333333334,
|
| 1822 |
+
"grad_norm": 0.2041015625,
|
| 1823 |
+
"kd_loss": 0.658003289926637,
|
| 1824 |
+
"learning_rate": 3e-06,
|
| 1825 |
+
"loss": 0.5988,
|
| 1826 |
+
"masked_tokens": 132.55,
|
| 1827 |
+
"mean_t": 0.5625698395539075,
|
| 1828 |
+
"step": 700,
|
| 1829 |
+
"student_masked_tokens": 132.55
|
| 1830 |
+
},
|
| 1831 |
+
{
|
| 1832 |
+
"avg_mask_ratio": 0.4790433386107907,
|
| 1833 |
+
"avg_response_length": 212.5,
|
| 1834 |
+
"avg_student_mask_ratio": 0.4790433386107907,
|
| 1835 |
+
"batch_ainp_frac": 0.0,
|
| 1836 |
+
"batch_inp_frac": 1.0,
|
| 1837 |
+
"batch_inp_oh_frac": 0.0,
|
| 1838 |
+
"batch_inp_par_frac": 0.0,
|
| 1839 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1840 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1841 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1842 |
+
"batch_rl_frac": 0.0,
|
| 1843 |
+
"batch_sft_frac": 0.0,
|
| 1844 |
+
"batch_soft_sft_frac": 0.0,
|
| 1845 |
+
"batch_tf_frac": 0.0,
|
| 1846 |
+
"ce_loss": 0.0,
|
| 1847 |
+
"epoch": 1.5162666666666667,
|
| 1848 |
+
"grad_norm": 0.173828125,
|
| 1849 |
+
"kd_loss": 0.42669102203761894,
|
| 1850 |
+
"learning_rate": 3e-06,
|
| 1851 |
+
"loss": 0.4103,
|
| 1852 |
+
"masked_tokens": 108.7375,
|
| 1853 |
+
"mean_t": 0.5135623761918395,
|
| 1854 |
+
"step": 710,
|
| 1855 |
+
"student_masked_tokens": 108.7375
|
| 1856 |
+
},
|
| 1857 |
+
{
|
| 1858 |
+
"avg_mask_ratio": 0.47950589570682495,
|
| 1859 |
+
"avg_response_length": 227.075,
|
| 1860 |
+
"avg_student_mask_ratio": 0.47950589570682495,
|
| 1861 |
+
"batch_ainp_frac": 0.0,
|
| 1862 |
+
"batch_inp_frac": 1.0,
|
| 1863 |
+
"batch_inp_oh_frac": 0.0,
|
| 1864 |
+
"batch_inp_par_frac": 0.0,
|
| 1865 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1866 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1867 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1868 |
+
"batch_rl_frac": 0.0,
|
| 1869 |
+
"batch_sft_frac": 0.0,
|
| 1870 |
+
"batch_soft_sft_frac": 0.0,
|
| 1871 |
+
"batch_tf_frac": 0.0,
|
| 1872 |
+
"ce_loss": 0.0,
|
| 1873 |
+
"epoch": 1.5375999999999999,
|
| 1874 |
+
"grad_norm": 0.365234375,
|
| 1875 |
+
"kd_loss": 0.562004916462638,
|
| 1876 |
+
"learning_rate": 3e-06,
|
| 1877 |
+
"loss": 0.5257,
|
| 1878 |
+
"masked_tokens": 110.8,
|
| 1879 |
+
"mean_t": 0.5117021896177902,
|
| 1880 |
+
"step": 720,
|
| 1881 |
+
"student_masked_tokens": 110.8
|
| 1882 |
+
},
|
| 1883 |
+
{
|
| 1884 |
+
"avg_mask_ratio": 0.4604924251558259,
|
| 1885 |
+
"avg_response_length": 232.925,
|
| 1886 |
+
"avg_student_mask_ratio": 0.4604924251558259,
|
| 1887 |
+
"batch_ainp_frac": 0.0,
|
| 1888 |
+
"batch_inp_frac": 1.0,
|
| 1889 |
+
"batch_inp_oh_frac": 0.0,
|
| 1890 |
+
"batch_inp_par_frac": 0.0,
|
| 1891 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1892 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1893 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1894 |
+
"batch_rl_frac": 0.0,
|
| 1895 |
+
"batch_sft_frac": 0.0,
|
| 1896 |
+
"batch_soft_sft_frac": 0.0,
|
| 1897 |
+
"batch_tf_frac": 0.0,
|
| 1898 |
+
"ce_loss": 0.0,
|
| 1899 |
+
"epoch": 1.5589333333333333,
|
| 1900 |
+
"grad_norm": 0.1630859375,
|
| 1901 |
+
"kd_loss": 0.42463611174693483,
|
| 1902 |
+
"learning_rate": 3e-06,
|
| 1903 |
+
"loss": 0.4182,
|
| 1904 |
+
"masked_tokens": 104.9625,
|
| 1905 |
+
"mean_t": 0.49050743713742123,
|
| 1906 |
+
"step": 730,
|
| 1907 |
+
"student_masked_tokens": 104.9625
|
| 1908 |
+
},
|
| 1909 |
+
{
|
| 1910 |
+
"avg_mask_ratio": 0.5185885130194947,
|
| 1911 |
+
"avg_response_length": 183.325,
|
| 1912 |
+
"avg_student_mask_ratio": 0.5185885130194947,
|
| 1913 |
+
"batch_ainp_frac": 0.0,
|
| 1914 |
+
"batch_inp_frac": 1.0,
|
| 1915 |
+
"batch_inp_oh_frac": 0.0,
|
| 1916 |
+
"batch_inp_par_frac": 0.0,
|
| 1917 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1918 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1919 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1920 |
+
"batch_rl_frac": 0.0,
|
| 1921 |
+
"batch_sft_frac": 0.0,
|
| 1922 |
+
"batch_soft_sft_frac": 0.0,
|
| 1923 |
+
"batch_tf_frac": 0.0,
|
| 1924 |
+
"ce_loss": 0.0,
|
| 1925 |
+
"epoch": 1.5802666666666667,
|
| 1926 |
+
"grad_norm": 0.1298828125,
|
| 1927 |
+
"kd_loss": 0.5289002722447321,
|
| 1928 |
+
"learning_rate": 3e-06,
|
| 1929 |
+
"loss": 0.5279,
|
| 1930 |
+
"masked_tokens": 97.125,
|
| 1931 |
+
"mean_t": 0.5505168779753149,
|
| 1932 |
+
"step": 740,
|
| 1933 |
+
"student_masked_tokens": 97.125
|
| 1934 |
+
},
|
| 1935 |
+
{
|
| 1936 |
+
"avg_mask_ratio": 0.4191439319110941,
|
| 1937 |
+
"avg_response_length": 223.65,
|
| 1938 |
+
"avg_student_mask_ratio": 0.4191439319110941,
|
| 1939 |
+
"batch_ainp_frac": 0.0,
|
| 1940 |
+
"batch_inp_frac": 1.0,
|
| 1941 |
+
"batch_inp_oh_frac": 0.0,
|
| 1942 |
+
"batch_inp_par_frac": 0.0,
|
| 1943 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1944 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1945 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1946 |
+
"batch_rl_frac": 0.0,
|
| 1947 |
+
"batch_sft_frac": 0.0,
|
| 1948 |
+
"batch_soft_sft_frac": 0.0,
|
| 1949 |
+
"batch_tf_frac": 0.0,
|
| 1950 |
+
"ce_loss": 0.0,
|
| 1951 |
+
"epoch": 1.6016,
|
| 1952 |
+
"grad_norm": 0.271484375,
|
| 1953 |
+
"kd_loss": 0.49676298911740346,
|
| 1954 |
+
"learning_rate": 3e-06,
|
| 1955 |
+
"loss": 0.4776,
|
| 1956 |
+
"masked_tokens": 95.3125,
|
| 1957 |
+
"mean_t": 0.4437690361432033,
|
| 1958 |
+
"step": 750,
|
| 1959 |
+
"student_masked_tokens": 95.3125
|
| 1960 |
+
},
|
| 1961 |
+
{
|
| 1962 |
+
"avg_mask_ratio": 0.46706983938929625,
|
| 1963 |
+
"avg_response_length": 216.0625,
|
| 1964 |
+
"avg_student_mask_ratio": 0.46706983938929625,
|
| 1965 |
+
"batch_ainp_frac": 0.0,
|
| 1966 |
+
"batch_inp_frac": 1.0,
|
| 1967 |
+
"batch_inp_oh_frac": 0.0,
|
| 1968 |
+
"batch_inp_par_frac": 0.0,
|
| 1969 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1970 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1971 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1972 |
+
"batch_rl_frac": 0.0,
|
| 1973 |
+
"batch_sft_frac": 0.0,
|
| 1974 |
+
"batch_soft_sft_frac": 0.0,
|
| 1975 |
+
"batch_tf_frac": 0.0,
|
| 1976 |
+
"ce_loss": 0.0,
|
| 1977 |
+
"epoch": 1.6229333333333333,
|
| 1978 |
+
"grad_norm": 1.234375,
|
| 1979 |
+
"kd_loss": 0.5423888156364214,
|
| 1980 |
+
"learning_rate": 3e-06,
|
| 1981 |
+
"loss": 0.5307,
|
| 1982 |
+
"masked_tokens": 107.7,
|
| 1983 |
+
"mean_t": 0.49132869170280175,
|
| 1984 |
+
"step": 760,
|
| 1985 |
+
"student_masked_tokens": 107.7
|
| 1986 |
+
},
|
| 1987 |
+
{
|
| 1988 |
+
"avg_mask_ratio": 0.44156218122225255,
|
| 1989 |
+
"avg_response_length": 259.675,
|
| 1990 |
+
"avg_student_mask_ratio": 0.44156218122225255,
|
| 1991 |
+
"batch_ainp_frac": 0.0,
|
| 1992 |
+
"batch_inp_frac": 1.0,
|
| 1993 |
+
"batch_inp_oh_frac": 0.0,
|
| 1994 |
+
"batch_inp_par_frac": 0.0,
|
| 1995 |
+
"batch_inp_par_par_frac": 0.0,
|
| 1996 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 1997 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 1998 |
+
"batch_rl_frac": 0.0,
|
| 1999 |
+
"batch_sft_frac": 0.0,
|
| 2000 |
+
"batch_soft_sft_frac": 0.0,
|
| 2001 |
+
"batch_tf_frac": 0.0,
|
| 2002 |
+
"ce_loss": 0.0,
|
| 2003 |
+
"epoch": 1.6442666666666668,
|
| 2004 |
+
"grad_norm": 0.1455078125,
|
| 2005 |
+
"kd_loss": 0.4092745722692257,
|
| 2006 |
+
"learning_rate": 3e-06,
|
| 2007 |
+
"loss": 0.3866,
|
| 2008 |
+
"masked_tokens": 117.5,
|
| 2009 |
+
"mean_t": 0.4667695587326307,
|
| 2010 |
+
"step": 770,
|
| 2011 |
+
"student_masked_tokens": 117.5
|
| 2012 |
+
},
|
| 2013 |
+
{
|
| 2014 |
+
"avg_mask_ratio": 0.42836043585848527,
|
| 2015 |
+
"avg_response_length": 258.5125,
|
| 2016 |
+
"avg_student_mask_ratio": 0.42836043585848527,
|
| 2017 |
+
"batch_ainp_frac": 0.0,
|
| 2018 |
+
"batch_inp_frac": 1.0,
|
| 2019 |
+
"batch_inp_oh_frac": 0.0,
|
| 2020 |
+
"batch_inp_par_frac": 0.0,
|
| 2021 |
+
"batch_inp_par_par_frac": 0.0,
|
| 2022 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 2023 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 2024 |
+
"batch_rl_frac": 0.0,
|
| 2025 |
+
"batch_sft_frac": 0.0,
|
| 2026 |
+
"batch_soft_sft_frac": 0.0,
|
| 2027 |
+
"batch_tf_frac": 0.0,
|
| 2028 |
+
"ce_loss": 0.0,
|
| 2029 |
+
"epoch": 1.6656,
|
| 2030 |
+
"grad_norm": 0.216796875,
|
| 2031 |
+
"kd_loss": 0.3423420360170297,
|
| 2032 |
+
"learning_rate": 3e-06,
|
| 2033 |
+
"loss": 0.3535,
|
| 2034 |
+
"masked_tokens": 119.0125,
|
| 2035 |
+
"mean_t": 0.44942845597106496,
|
| 2036 |
+
"step": 780,
|
| 2037 |
+
"student_masked_tokens": 119.0125
|
| 2038 |
+
},
|
| 2039 |
+
{
|
| 2040 |
+
"avg_mask_ratio": 0.46589430308085866,
|
| 2041 |
+
"avg_response_length": 222.3125,
|
| 2042 |
+
"avg_student_mask_ratio": 0.46589430308085866,
|
| 2043 |
+
"batch_ainp_frac": 0.0,
|
| 2044 |
+
"batch_inp_frac": 1.0,
|
| 2045 |
+
"batch_inp_oh_frac": 0.0,
|
| 2046 |
+
"batch_inp_par_frac": 0.0,
|
| 2047 |
+
"batch_inp_par_par_frac": 0.0,
|
| 2048 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 2049 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 2050 |
+
"batch_rl_frac": 0.0,
|
| 2051 |
+
"batch_sft_frac": 0.0,
|
| 2052 |
+
"batch_soft_sft_frac": 0.0,
|
| 2053 |
+
"batch_tf_frac": 0.0,
|
| 2054 |
+
"ce_loss": 0.0,
|
| 2055 |
+
"epoch": 1.6869333333333332,
|
| 2056 |
+
"grad_norm": 0.125,
|
| 2057 |
+
"kd_loss": 0.33727191750285523,
|
| 2058 |
+
"learning_rate": 3e-06,
|
| 2059 |
+
"loss": 0.3767,
|
| 2060 |
+
"masked_tokens": 103.25,
|
| 2061 |
+
"mean_t": 0.4993515375303105,
|
| 2062 |
+
"step": 790,
|
| 2063 |
+
"student_masked_tokens": 103.25
|
| 2064 |
+
},
|
| 2065 |
+
{
|
| 2066 |
+
"avg_mask_ratio": 0.46366424662992356,
|
| 2067 |
+
"avg_response_length": 219.6875,
|
| 2068 |
+
"avg_student_mask_ratio": 0.46366424662992356,
|
| 2069 |
+
"batch_ainp_frac": 0.0,
|
| 2070 |
+
"batch_inp_frac": 1.0,
|
| 2071 |
+
"batch_inp_oh_frac": 0.0,
|
| 2072 |
+
"batch_inp_par_frac": 0.0,
|
| 2073 |
+
"batch_inp_par_par_frac": 0.0,
|
| 2074 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 2075 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 2076 |
+
"batch_rl_frac": 0.0,
|
| 2077 |
+
"batch_sft_frac": 0.0,
|
| 2078 |
+
"batch_soft_sft_frac": 0.0,
|
| 2079 |
+
"batch_tf_frac": 0.0,
|
| 2080 |
+
"ce_loss": 0.0,
|
| 2081 |
+
"epoch": 1.7082666666666668,
|
| 2082 |
+
"grad_norm": 0.189453125,
|
| 2083 |
+
"kd_loss": 0.34817896180302343,
|
| 2084 |
+
"learning_rate": 3e-06,
|
| 2085 |
+
"loss": 0.3863,
|
| 2086 |
+
"masked_tokens": 104.5,
|
| 2087 |
+
"mean_t": 0.500370389316231,
|
| 2088 |
+
"step": 800,
|
| 2089 |
+
"student_masked_tokens": 104.5
|
| 2090 |
+
},
|
| 2091 |
+
{
|
| 2092 |
+
"avg_mask_ratio": 0.503375941584818,
|
| 2093 |
+
"avg_response_length": 237.85,
|
| 2094 |
+
"avg_student_mask_ratio": 0.503375941584818,
|
| 2095 |
+
"batch_ainp_frac": 0.0,
|
| 2096 |
+
"batch_inp_frac": 1.0,
|
| 2097 |
+
"batch_inp_oh_frac": 0.0,
|
| 2098 |
+
"batch_inp_par_frac": 0.0,
|
| 2099 |
+
"batch_inp_par_par_frac": 0.0,
|
| 2100 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 2101 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 2102 |
+
"batch_rl_frac": 0.0,
|
| 2103 |
+
"batch_sft_frac": 0.0,
|
| 2104 |
+
"batch_soft_sft_frac": 0.0,
|
| 2105 |
+
"batch_tf_frac": 0.0,
|
| 2106 |
+
"ce_loss": 0.0,
|
| 2107 |
+
"epoch": 1.7296,
|
| 2108 |
+
"grad_norm": 0.234375,
|
| 2109 |
+
"kd_loss": 0.4517137570231853,
|
| 2110 |
+
"learning_rate": 3e-06,
|
| 2111 |
+
"loss": 0.464,
|
| 2112 |
+
"masked_tokens": 118.2125,
|
| 2113 |
+
"mean_t": 0.5317009104182944,
|
| 2114 |
+
"step": 810,
|
| 2115 |
+
"student_masked_tokens": 118.2125
|
| 2116 |
+
},
|
| 2117 |
+
{
|
| 2118 |
+
"avg_mask_ratio": 0.5110091455746442,
|
| 2119 |
+
"avg_response_length": 209.0875,
|
| 2120 |
+
"avg_student_mask_ratio": 0.5110091455746442,
|
| 2121 |
+
"batch_ainp_frac": 0.0,
|
| 2122 |
+
"batch_inp_frac": 1.0,
|
| 2123 |
+
"batch_inp_oh_frac": 0.0,
|
| 2124 |
+
"batch_inp_par_frac": 0.0,
|
| 2125 |
+
"batch_inp_par_par_frac": 0.0,
|
| 2126 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 2127 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 2128 |
+
"batch_rl_frac": 0.0,
|
| 2129 |
+
"batch_sft_frac": 0.0,
|
| 2130 |
+
"batch_soft_sft_frac": 0.0,
|
| 2131 |
+
"batch_tf_frac": 0.0,
|
| 2132 |
+
"ce_loss": 0.0,
|
| 2133 |
+
"epoch": 1.7509333333333332,
|
| 2134 |
+
"grad_norm": 0.33203125,
|
| 2135 |
+
"kd_loss": 0.590595377480895,
|
| 2136 |
+
"learning_rate": 3e-06,
|
| 2137 |
+
"loss": 0.5476,
|
| 2138 |
+
"masked_tokens": 120.95,
|
| 2139 |
+
"mean_t": 0.5392061032878701,
|
| 2140 |
+
"step": 820,
|
| 2141 |
+
"student_masked_tokens": 120.95
|
| 2142 |
+
},
|
| 2143 |
+
{
|
| 2144 |
+
"avg_mask_ratio": 0.49899387182667854,
|
| 2145 |
+
"avg_response_length": 263.975,
|
| 2146 |
+
"avg_student_mask_ratio": 0.49899387182667854,
|
| 2147 |
+
"batch_ainp_frac": 0.0,
|
| 2148 |
+
"batch_inp_frac": 1.0,
|
| 2149 |
+
"batch_inp_oh_frac": 0.0,
|
| 2150 |
+
"batch_inp_par_frac": 0.0,
|
| 2151 |
+
"batch_inp_par_par_frac": 0.0,
|
| 2152 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 2153 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 2154 |
+
"batch_rl_frac": 0.0,
|
| 2155 |
+
"batch_sft_frac": 0.0,
|
| 2156 |
+
"batch_soft_sft_frac": 0.0,
|
| 2157 |
+
"batch_tf_frac": 0.0,
|
| 2158 |
+
"ce_loss": 0.0,
|
| 2159 |
+
"epoch": 1.7722666666666667,
|
| 2160 |
+
"grad_norm": 0.1455078125,
|
| 2161 |
+
"kd_loss": 0.5558899336714603,
|
| 2162 |
+
"learning_rate": 3e-06,
|
| 2163 |
+
"loss": 0.5794,
|
| 2164 |
+
"masked_tokens": 137.075,
|
| 2165 |
+
"mean_t": 0.5238314627087675,
|
| 2166 |
+
"step": 830,
|
| 2167 |
+
"student_masked_tokens": 137.075
|
| 2168 |
+
},
|
| 2169 |
+
{
|
| 2170 |
+
"avg_mask_ratio": 0.4997270987310912,
|
| 2171 |
+
"avg_response_length": 221.9,
|
| 2172 |
+
"avg_student_mask_ratio": 0.4997270987310912,
|
| 2173 |
+
"batch_ainp_frac": 0.0,
|
| 2174 |
+
"batch_inp_frac": 1.0,
|
| 2175 |
+
"batch_inp_oh_frac": 0.0,
|
| 2176 |
+
"batch_inp_par_frac": 0.0,
|
| 2177 |
+
"batch_inp_par_par_frac": 0.0,
|
| 2178 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 2179 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 2180 |
+
"batch_rl_frac": 0.0,
|
| 2181 |
+
"batch_sft_frac": 0.0,
|
| 2182 |
+
"batch_soft_sft_frac": 0.0,
|
| 2183 |
+
"batch_tf_frac": 0.0,
|
| 2184 |
+
"ce_loss": 0.0,
|
| 2185 |
+
"epoch": 1.7936,
|
| 2186 |
+
"grad_norm": 0.126953125,
|
| 2187 |
+
"kd_loss": 0.4676562640790394,
|
| 2188 |
+
"learning_rate": 3e-06,
|
| 2189 |
+
"loss": 0.4654,
|
| 2190 |
+
"masked_tokens": 114.525,
|
| 2191 |
+
"mean_t": 0.5301066277665086,
|
| 2192 |
+
"step": 840,
|
| 2193 |
+
"student_masked_tokens": 114.525
|
| 2194 |
+
},
|
| 2195 |
+
{
|
| 2196 |
+
"avg_mask_ratio": 0.4988076956477016,
|
| 2197 |
+
"avg_response_length": 225.5,
|
| 2198 |
+
"avg_student_mask_ratio": 0.4988076956477016,
|
| 2199 |
+
"batch_ainp_frac": 0.0,
|
| 2200 |
+
"batch_inp_frac": 1.0,
|
| 2201 |
+
"batch_inp_oh_frac": 0.0,
|
| 2202 |
+
"batch_inp_par_frac": 0.0,
|
| 2203 |
+
"batch_inp_par_par_frac": 0.0,
|
| 2204 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 2205 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 2206 |
+
"batch_rl_frac": 0.0,
|
| 2207 |
+
"batch_sft_frac": 0.0,
|
| 2208 |
+
"batch_soft_sft_frac": 0.0,
|
| 2209 |
+
"batch_tf_frac": 0.0,
|
| 2210 |
+
"ce_loss": 0.0,
|
| 2211 |
+
"epoch": 1.8149333333333333,
|
| 2212 |
+
"grad_norm": 0.11474609375,
|
| 2213 |
+
"kd_loss": 0.3244839489429069,
|
| 2214 |
+
"learning_rate": 3e-06,
|
| 2215 |
+
"loss": 0.3941,
|
| 2216 |
+
"masked_tokens": 116.9125,
|
| 2217 |
+
"mean_t": 0.5343429344706238,
|
| 2218 |
+
"step": 850,
|
| 2219 |
+
"student_masked_tokens": 116.9125
|
| 2220 |
+
},
|
| 2221 |
+
{
|
| 2222 |
+
"avg_mask_ratio": 0.4635998342186213,
|
| 2223 |
+
"avg_response_length": 229.125,
|
| 2224 |
+
"avg_student_mask_ratio": 0.4635998342186213,
|
| 2225 |
+
"batch_ainp_frac": 0.0,
|
| 2226 |
+
"batch_inp_frac": 1.0,
|
| 2227 |
+
"batch_inp_oh_frac": 0.0,
|
| 2228 |
+
"batch_inp_par_frac": 0.0,
|
| 2229 |
+
"batch_inp_par_par_frac": 0.0,
|
| 2230 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 2231 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 2232 |
+
"batch_rl_frac": 0.0,
|
| 2233 |
+
"batch_sft_frac": 0.0,
|
| 2234 |
+
"batch_soft_sft_frac": 0.0,
|
| 2235 |
+
"batch_tf_frac": 0.0,
|
| 2236 |
+
"ce_loss": 0.0,
|
| 2237 |
+
"epoch": 1.8362666666666667,
|
| 2238 |
+
"grad_norm": 0.181640625,
|
| 2239 |
+
"kd_loss": 0.4306980623789968,
|
| 2240 |
+
"learning_rate": 3e-06,
|
| 2241 |
+
"loss": 0.4216,
|
| 2242 |
+
"masked_tokens": 109.675,
|
| 2243 |
+
"mean_t": 0.4791536889737472,
|
| 2244 |
+
"step": 860,
|
| 2245 |
+
"student_masked_tokens": 109.675
|
| 2246 |
+
},
|
| 2247 |
+
{
|
| 2248 |
+
"avg_mask_ratio": 0.49111039767740294,
|
| 2249 |
+
"avg_response_length": 229.1,
|
| 2250 |
+
"avg_student_mask_ratio": 0.49111039767740294,
|
| 2251 |
+
"batch_ainp_frac": 0.0,
|
| 2252 |
+
"batch_inp_frac": 1.0,
|
| 2253 |
+
"batch_inp_oh_frac": 0.0,
|
| 2254 |
+
"batch_inp_par_frac": 0.0,
|
| 2255 |
+
"batch_inp_par_par_frac": 0.0,
|
| 2256 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 2257 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 2258 |
+
"batch_rl_frac": 0.0,
|
| 2259 |
+
"batch_sft_frac": 0.0,
|
| 2260 |
+
"batch_soft_sft_frac": 0.0,
|
| 2261 |
+
"batch_tf_frac": 0.0,
|
| 2262 |
+
"ce_loss": 0.0,
|
| 2263 |
+
"epoch": 1.8576000000000001,
|
| 2264 |
+
"grad_norm": 0.2412109375,
|
| 2265 |
+
"kd_loss": 0.4694033406512972,
|
| 2266 |
+
"learning_rate": 3e-06,
|
| 2267 |
+
"loss": 0.5022,
|
| 2268 |
+
"masked_tokens": 115.6625,
|
| 2269 |
+
"mean_t": 0.5203817339061061,
|
| 2270 |
+
"step": 870,
|
| 2271 |
+
"student_masked_tokens": 115.6625
|
| 2272 |
+
},
|
| 2273 |
+
{
|
| 2274 |
+
"avg_mask_ratio": 0.4427660425659269,
|
| 2275 |
+
"avg_response_length": 198.5625,
|
| 2276 |
+
"avg_student_mask_ratio": 0.4427660425659269,
|
| 2277 |
+
"batch_ainp_frac": 0.0,
|
| 2278 |
+
"batch_inp_frac": 1.0,
|
| 2279 |
+
"batch_inp_oh_frac": 0.0,
|
| 2280 |
+
"batch_inp_par_frac": 0.0,
|
| 2281 |
+
"batch_inp_par_par_frac": 0.0,
|
| 2282 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 2283 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 2284 |
+
"batch_rl_frac": 0.0,
|
| 2285 |
+
"batch_sft_frac": 0.0,
|
| 2286 |
+
"batch_soft_sft_frac": 0.0,
|
| 2287 |
+
"batch_tf_frac": 0.0,
|
| 2288 |
+
"ce_loss": 0.0,
|
| 2289 |
+
"epoch": 1.8789333333333333,
|
| 2290 |
+
"grad_norm": 0.1728515625,
|
| 2291 |
+
"kd_loss": 0.46008693509684234,
|
| 2292 |
+
"learning_rate": 3e-06,
|
| 2293 |
+
"loss": 0.4582,
|
| 2294 |
+
"masked_tokens": 91.175,
|
| 2295 |
+
"mean_t": 0.4875184997683391,
|
| 2296 |
+
"step": 880,
|
| 2297 |
+
"student_masked_tokens": 91.175
|
| 2298 |
+
},
|
| 2299 |
+
{
|
| 2300 |
+
"avg_mask_ratio": 0.4464349385118112,
|
| 2301 |
+
"avg_response_length": 225.8375,
|
| 2302 |
+
"avg_student_mask_ratio": 0.4464349385118112,
|
| 2303 |
+
"batch_ainp_frac": 0.0,
|
| 2304 |
+
"batch_inp_frac": 1.0,
|
| 2305 |
+
"batch_inp_oh_frac": 0.0,
|
| 2306 |
+
"batch_inp_par_frac": 0.0,
|
| 2307 |
+
"batch_inp_par_par_frac": 0.0,
|
| 2308 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 2309 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 2310 |
+
"batch_rl_frac": 0.0,
|
| 2311 |
+
"batch_sft_frac": 0.0,
|
| 2312 |
+
"batch_soft_sft_frac": 0.0,
|
| 2313 |
+
"batch_tf_frac": 0.0,
|
| 2314 |
+
"ce_loss": 0.0,
|
| 2315 |
+
"epoch": 1.9002666666666665,
|
| 2316 |
+
"grad_norm": 0.0888671875,
|
| 2317 |
+
"kd_loss": 0.3971562616849951,
|
| 2318 |
+
"learning_rate": 3e-06,
|
| 2319 |
+
"loss": 0.3817,
|
| 2320 |
+
"masked_tokens": 101.75,
|
| 2321 |
+
"mean_t": 0.4766692223958671,
|
| 2322 |
+
"step": 890,
|
| 2323 |
+
"student_masked_tokens": 101.75
|
| 2324 |
+
},
|
| 2325 |
+
{
|
| 2326 |
+
"avg_mask_ratio": 0.44976164362160487,
|
| 2327 |
+
"avg_response_length": 227.7875,
|
| 2328 |
+
"avg_student_mask_ratio": 0.44976164362160487,
|
| 2329 |
+
"batch_ainp_frac": 0.0,
|
| 2330 |
+
"batch_inp_frac": 1.0,
|
| 2331 |
+
"batch_inp_oh_frac": 0.0,
|
| 2332 |
+
"batch_inp_par_frac": 0.0,
|
| 2333 |
+
"batch_inp_par_par_frac": 0.0,
|
| 2334 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 2335 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 2336 |
+
"batch_rl_frac": 0.0,
|
| 2337 |
+
"batch_sft_frac": 0.0,
|
| 2338 |
+
"batch_soft_sft_frac": 0.0,
|
| 2339 |
+
"batch_tf_frac": 0.0,
|
| 2340 |
+
"ce_loss": 0.0,
|
| 2341 |
+
"epoch": 1.9216,
|
| 2342 |
+
"grad_norm": 0.1708984375,
|
| 2343 |
+
"kd_loss": 0.4591321104033284,
|
| 2344 |
+
"learning_rate": 3e-06,
|
| 2345 |
+
"loss": 0.4498,
|
| 2346 |
+
"masked_tokens": 103.475,
|
| 2347 |
+
"mean_t": 0.487134758150205,
|
| 2348 |
+
"step": 900,
|
| 2349 |
+
"student_masked_tokens": 103.475
|
| 2350 |
+
},
|
| 2351 |
+
{
|
| 2352 |
+
"avg_mask_ratio": 0.475579984736396,
|
| 2353 |
+
"avg_response_length": 245.1625,
|
| 2354 |
+
"avg_student_mask_ratio": 0.475579984736396,
|
| 2355 |
+
"batch_ainp_frac": 0.0,
|
| 2356 |
+
"batch_inp_frac": 1.0,
|
| 2357 |
+
"batch_inp_oh_frac": 0.0,
|
| 2358 |
+
"batch_inp_par_frac": 0.0,
|
| 2359 |
+
"batch_inp_par_par_frac": 0.0,
|
| 2360 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 2361 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 2362 |
+
"batch_rl_frac": 0.0,
|
| 2363 |
+
"batch_sft_frac": 0.0,
|
| 2364 |
+
"batch_soft_sft_frac": 0.0,
|
| 2365 |
+
"batch_tf_frac": 0.0,
|
| 2366 |
+
"ce_loss": 0.0,
|
| 2367 |
+
"epoch": 1.9429333333333334,
|
| 2368 |
+
"grad_norm": 0.1328125,
|
| 2369 |
+
"kd_loss": 0.4644632683706561,
|
| 2370 |
+
"learning_rate": 3e-06,
|
| 2371 |
+
"loss": 0.4135,
|
| 2372 |
+
"masked_tokens": 124.6375,
|
| 2373 |
+
"mean_t": 0.5027793228859082,
|
| 2374 |
+
"step": 910,
|
| 2375 |
+
"student_masked_tokens": 124.6375
|
| 2376 |
+
},
|
| 2377 |
+
{
|
| 2378 |
+
"avg_mask_ratio": 0.4688875659601763,
|
| 2379 |
+
"avg_response_length": 226.2875,
|
| 2380 |
+
"avg_student_mask_ratio": 0.4688875659601763,
|
| 2381 |
+
"batch_ainp_frac": 0.0,
|
| 2382 |
+
"batch_inp_frac": 1.0,
|
| 2383 |
+
"batch_inp_oh_frac": 0.0,
|
| 2384 |
+
"batch_inp_par_frac": 0.0,
|
| 2385 |
+
"batch_inp_par_par_frac": 0.0,
|
| 2386 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 2387 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 2388 |
+
"batch_rl_frac": 0.0,
|
| 2389 |
+
"batch_sft_frac": 0.0,
|
| 2390 |
+
"batch_soft_sft_frac": 0.0,
|
| 2391 |
+
"batch_tf_frac": 0.0,
|
| 2392 |
+
"ce_loss": 0.0,
|
| 2393 |
+
"epoch": 1.9642666666666666,
|
| 2394 |
+
"grad_norm": 0.2099609375,
|
| 2395 |
+
"kd_loss": 0.4406846639433752,
|
| 2396 |
+
"learning_rate": 3e-06,
|
| 2397 |
+
"loss": 0.4429,
|
| 2398 |
+
"masked_tokens": 110.0125,
|
| 2399 |
+
"mean_t": 0.49417946098838,
|
| 2400 |
+
"step": 920,
|
| 2401 |
+
"student_masked_tokens": 110.0125
|
| 2402 |
+
},
|
| 2403 |
+
{
|
| 2404 |
+
"avg_mask_ratio": 0.47999348094454036,
|
| 2405 |
+
"avg_response_length": 237.05,
|
| 2406 |
+
"avg_student_mask_ratio": 0.47999348094454036,
|
| 2407 |
+
"batch_ainp_frac": 0.0,
|
| 2408 |
+
"batch_inp_frac": 1.0,
|
| 2409 |
+
"batch_inp_oh_frac": 0.0,
|
| 2410 |
+
"batch_inp_par_frac": 0.0,
|
| 2411 |
+
"batch_inp_par_par_frac": 0.0,
|
| 2412 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 2413 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 2414 |
+
"batch_rl_frac": 0.0,
|
| 2415 |
+
"batch_sft_frac": 0.0,
|
| 2416 |
+
"batch_soft_sft_frac": 0.0,
|
| 2417 |
+
"batch_tf_frac": 0.0,
|
| 2418 |
+
"ce_loss": 0.0,
|
| 2419 |
+
"epoch": 1.9856,
|
| 2420 |
+
"grad_norm": 0.2275390625,
|
| 2421 |
+
"kd_loss": 0.4538009871917458,
|
| 2422 |
+
"learning_rate": 3e-06,
|
| 2423 |
+
"loss": 0.4455,
|
| 2424 |
+
"masked_tokens": 121.6,
|
| 2425 |
+
"mean_t": 0.5045580042526125,
|
| 2426 |
+
"step": 930,
|
| 2427 |
+
"student_masked_tokens": 121.6
|
| 2428 |
+
},
|
| 2429 |
+
{
|
| 2430 |
+
"avg_mask_ratio": 0.49413903727240505,
|
| 2431 |
+
"avg_response_length": 224.79761904761904,
|
| 2432 |
+
"avg_student_mask_ratio": 0.49413903727240505,
|
| 2433 |
+
"batch_ainp_frac": 0.0,
|
| 2434 |
+
"batch_inp_frac": 1.0,
|
| 2435 |
+
"batch_inp_oh_frac": 0.0,
|
| 2436 |
+
"batch_inp_par_frac": 0.0,
|
| 2437 |
+
"batch_inp_par_par_frac": 0.0,
|
| 2438 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 2439 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 2440 |
+
"batch_rl_frac": 0.0,
|
| 2441 |
+
"batch_sft_frac": 0.0,
|
| 2442 |
+
"batch_soft_sft_frac": 0.0,
|
| 2443 |
+
"batch_tf_frac": 0.0,
|
| 2444 |
+
"ce_loss": 0.0,
|
| 2445 |
+
"epoch": 2.0085333333333333,
|
| 2446 |
+
"grad_norm": 0.294921875,
|
| 2447 |
+
"kd_loss": 0.4914910109509676,
|
| 2448 |
+
"learning_rate": 3e-06,
|
| 2449 |
+
"loss": 0.4787,
|
| 2450 |
+
"masked_tokens": 120.5,
|
| 2451 |
+
"mean_t": 0.5321138524893849,
|
| 2452 |
+
"step": 940,
|
| 2453 |
+
"student_masked_tokens": 120.5
|
| 2454 |
+
},
|
| 2455 |
+
{
|
| 2456 |
+
"avg_mask_ratio": 0.4368605303927325,
|
| 2457 |
+
"avg_response_length": 240.9125,
|
| 2458 |
+
"avg_student_mask_ratio": 0.4368605303927325,
|
| 2459 |
+
"batch_ainp_frac": 0.0,
|
| 2460 |
+
"batch_inp_frac": 1.0,
|
| 2461 |
+
"batch_inp_oh_frac": 0.0,
|
| 2462 |
+
"batch_inp_par_frac": 0.0,
|
| 2463 |
+
"batch_inp_par_par_frac": 0.0,
|
| 2464 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 2465 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 2466 |
+
"batch_rl_frac": 0.0,
|
| 2467 |
+
"batch_sft_frac": 0.0,
|
| 2468 |
+
"batch_soft_sft_frac": 0.0,
|
| 2469 |
+
"batch_tf_frac": 0.0,
|
| 2470 |
+
"ce_loss": 0.0,
|
| 2471 |
+
"epoch": 2.0298666666666665,
|
| 2472 |
+
"grad_norm": 0.162109375,
|
| 2473 |
+
"kd_loss": 0.42506327110539815,
|
| 2474 |
+
"learning_rate": 3e-06,
|
| 2475 |
+
"loss": 0.4051,
|
| 2476 |
+
"masked_tokens": 111.4125,
|
| 2477 |
+
"mean_t": 0.4632946296595037,
|
| 2478 |
+
"step": 950,
|
| 2479 |
+
"student_masked_tokens": 111.4125
|
| 2480 |
+
},
|
| 2481 |
+
{
|
| 2482 |
+
"avg_mask_ratio": 0.4988762516761199,
|
| 2483 |
+
"avg_response_length": 275.3,
|
| 2484 |
+
"avg_student_mask_ratio": 0.4988762516761199,
|
| 2485 |
+
"batch_ainp_frac": 0.0,
|
| 2486 |
+
"batch_inp_frac": 1.0,
|
| 2487 |
+
"batch_inp_oh_frac": 0.0,
|
| 2488 |
+
"batch_inp_par_frac": 0.0,
|
| 2489 |
+
"batch_inp_par_par_frac": 0.0,
|
| 2490 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 2491 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 2492 |
+
"batch_rl_frac": 0.0,
|
| 2493 |
+
"batch_sft_frac": 0.0,
|
| 2494 |
+
"batch_soft_sft_frac": 0.0,
|
| 2495 |
+
"batch_tf_frac": 0.0,
|
| 2496 |
+
"ce_loss": 0.0,
|
| 2497 |
+
"epoch": 2.0512,
|
| 2498 |
+
"grad_norm": 0.1103515625,
|
| 2499 |
+
"kd_loss": 0.5790403748234894,
|
| 2500 |
+
"learning_rate": 3e-06,
|
| 2501 |
+
"loss": 0.5466,
|
| 2502 |
+
"masked_tokens": 143.825,
|
| 2503 |
+
"mean_t": 0.5198000721400604,
|
| 2504 |
+
"step": 960,
|
| 2505 |
+
"student_masked_tokens": 143.825
|
| 2506 |
+
},
|
| 2507 |
+
{
|
| 2508 |
+
"avg_mask_ratio": 0.437801384011982,
|
| 2509 |
+
"avg_response_length": 236.2375,
|
| 2510 |
+
"avg_student_mask_ratio": 0.437801384011982,
|
| 2511 |
+
"batch_ainp_frac": 0.0,
|
| 2512 |
+
"batch_inp_frac": 1.0,
|
| 2513 |
+
"batch_inp_oh_frac": 0.0,
|
| 2514 |
+
"batch_inp_par_frac": 0.0,
|
| 2515 |
+
"batch_inp_par_par_frac": 0.0,
|
| 2516 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 2517 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 2518 |
+
"batch_rl_frac": 0.0,
|
| 2519 |
+
"batch_sft_frac": 0.0,
|
| 2520 |
+
"batch_soft_sft_frac": 0.0,
|
| 2521 |
+
"batch_tf_frac": 0.0,
|
| 2522 |
+
"ce_loss": 0.0,
|
| 2523 |
+
"epoch": 2.0725333333333333,
|
| 2524 |
+
"grad_norm": 0.248046875,
|
| 2525 |
+
"kd_loss": 0.35248266551700097,
|
| 2526 |
+
"learning_rate": 3e-06,
|
| 2527 |
+
"loss": 0.3947,
|
| 2528 |
+
"masked_tokens": 107.025,
|
| 2529 |
+
"mean_t": 0.4703940597362816,
|
| 2530 |
+
"step": 970,
|
| 2531 |
+
"student_masked_tokens": 107.025
|
| 2532 |
+
},
|
| 2533 |
+
{
|
| 2534 |
+
"avg_mask_ratio": 0.42220073882490394,
|
| 2535 |
+
"avg_response_length": 230.8625,
|
| 2536 |
+
"avg_student_mask_ratio": 0.42220073882490394,
|
| 2537 |
+
"batch_ainp_frac": 0.0,
|
| 2538 |
+
"batch_inp_frac": 1.0,
|
| 2539 |
+
"batch_inp_oh_frac": 0.0,
|
| 2540 |
+
"batch_inp_par_frac": 0.0,
|
| 2541 |
+
"batch_inp_par_par_frac": 0.0,
|
| 2542 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 2543 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 2544 |
+
"batch_rl_frac": 0.0,
|
| 2545 |
+
"batch_sft_frac": 0.0,
|
| 2546 |
+
"batch_soft_sft_frac": 0.0,
|
| 2547 |
+
"batch_tf_frac": 0.0,
|
| 2548 |
+
"ce_loss": 0.0,
|
| 2549 |
+
"epoch": 2.0938666666666665,
|
| 2550 |
+
"grad_norm": 0.1943359375,
|
| 2551 |
+
"kd_loss": 0.3541594335687478,
|
| 2552 |
+
"learning_rate": 3e-06,
|
| 2553 |
+
"loss": 0.3645,
|
| 2554 |
+
"masked_tokens": 97.7875,
|
| 2555 |
+
"mean_t": 0.4511947895749472,
|
| 2556 |
+
"step": 980,
|
| 2557 |
+
"student_masked_tokens": 97.7875
|
| 2558 |
+
},
|
| 2559 |
+
{
|
| 2560 |
+
"avg_mask_ratio": 0.4605769342277199,
|
| 2561 |
+
"avg_response_length": 262.0375,
|
| 2562 |
+
"avg_student_mask_ratio": 0.4605769342277199,
|
| 2563 |
+
"batch_ainp_frac": 0.0,
|
| 2564 |
+
"batch_inp_frac": 1.0,
|
| 2565 |
+
"batch_inp_oh_frac": 0.0,
|
| 2566 |
+
"batch_inp_par_frac": 0.0,
|
| 2567 |
+
"batch_inp_par_par_frac": 0.0,
|
| 2568 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 2569 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 2570 |
+
"batch_rl_frac": 0.0,
|
| 2571 |
+
"batch_sft_frac": 0.0,
|
| 2572 |
+
"batch_soft_sft_frac": 0.0,
|
| 2573 |
+
"batch_tf_frac": 0.0,
|
| 2574 |
+
"ce_loss": 0.0,
|
| 2575 |
+
"epoch": 2.1152,
|
| 2576 |
+
"grad_norm": 0.2216796875,
|
| 2577 |
+
"kd_loss": 0.41152406695375704,
|
| 2578 |
+
"learning_rate": 3e-06,
|
| 2579 |
+
"loss": 0.4128,
|
| 2580 |
+
"masked_tokens": 122.0,
|
| 2581 |
+
"mean_t": 0.4923786667350214,
|
| 2582 |
+
"step": 990,
|
| 2583 |
+
"student_masked_tokens": 122.0
|
| 2584 |
+
},
|
| 2585 |
+
{
|
| 2586 |
+
"avg_mask_ratio": 0.4547682981239632,
|
| 2587 |
+
"avg_response_length": 215.3,
|
| 2588 |
+
"avg_student_mask_ratio": 0.4547682981239632,
|
| 2589 |
+
"batch_ainp_frac": 0.0,
|
| 2590 |
+
"batch_inp_frac": 1.0,
|
| 2591 |
+
"batch_inp_oh_frac": 0.0,
|
| 2592 |
+
"batch_inp_par_frac": 0.0,
|
| 2593 |
+
"batch_inp_par_par_frac": 0.0,
|
| 2594 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 2595 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 2596 |
+
"batch_rl_frac": 0.0,
|
| 2597 |
+
"batch_sft_frac": 0.0,
|
| 2598 |
+
"batch_soft_sft_frac": 0.0,
|
| 2599 |
+
"batch_tf_frac": 0.0,
|
| 2600 |
+
"ce_loss": 0.0,
|
| 2601 |
+
"epoch": 2.1365333333333334,
|
| 2602 |
+
"grad_norm": 0.326171875,
|
| 2603 |
+
"kd_loss": 0.3404828482017706,
|
| 2604 |
+
"learning_rate": 3e-06,
|
| 2605 |
+
"loss": 0.3793,
|
| 2606 |
+
"masked_tokens": 100.775,
|
| 2607 |
+
"mean_t": 0.4773523230338469,
|
| 2608 |
+
"step": 1000,
|
| 2609 |
+
"student_masked_tokens": 100.775
|
| 2610 |
+
},
|
| 2611 |
+
{
|
| 2612 |
+
"avg_mask_ratio": 0.43540415074676275,
|
| 2613 |
+
"avg_response_length": 215.0,
|
| 2614 |
+
"avg_student_mask_ratio": 0.43540415074676275,
|
| 2615 |
+
"batch_ainp_frac": 0.0,
|
| 2616 |
+
"batch_inp_frac": 1.0,
|
| 2617 |
+
"batch_inp_oh_frac": 0.0,
|
| 2618 |
+
"batch_inp_par_frac": 0.0,
|
| 2619 |
+
"batch_inp_par_par_frac": 0.0,
|
| 2620 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 2621 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 2622 |
+
"batch_rl_frac": 0.0,
|
| 2623 |
+
"batch_sft_frac": 0.0,
|
| 2624 |
+
"batch_soft_sft_frac": 0.0,
|
| 2625 |
+
"batch_tf_frac": 0.0,
|
| 2626 |
+
"ce_loss": 0.0,
|
| 2627 |
+
"epoch": 2.1578666666666666,
|
| 2628 |
+
"grad_norm": 0.2216796875,
|
| 2629 |
+
"kd_loss": 0.35834208256953526,
|
| 2630 |
+
"learning_rate": 3e-06,
|
| 2631 |
+
"loss": 0.3483,
|
| 2632 |
+
"masked_tokens": 88.65,
|
| 2633 |
+
"mean_t": 0.4648138735938119,
|
| 2634 |
+
"step": 1010,
|
| 2635 |
+
"student_masked_tokens": 88.65
|
| 2636 |
+
},
|
| 2637 |
+
{
|
| 2638 |
+
"avg_mask_ratio": 0.5063220548443497,
|
| 2639 |
+
"avg_response_length": 206.9125,
|
| 2640 |
+
"avg_student_mask_ratio": 0.5063220548443497,
|
| 2641 |
+
"batch_ainp_frac": 0.0,
|
| 2642 |
+
"batch_inp_frac": 1.0,
|
| 2643 |
+
"batch_inp_oh_frac": 0.0,
|
| 2644 |
+
"batch_inp_par_frac": 0.0,
|
| 2645 |
+
"batch_inp_par_par_frac": 0.0,
|
| 2646 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 2647 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 2648 |
+
"batch_rl_frac": 0.0,
|
| 2649 |
+
"batch_sft_frac": 0.0,
|
| 2650 |
+
"batch_soft_sft_frac": 0.0,
|
| 2651 |
+
"batch_tf_frac": 0.0,
|
| 2652 |
+
"ce_loss": 0.0,
|
| 2653 |
+
"epoch": 2.1792,
|
| 2654 |
+
"grad_norm": 0.080078125,
|
| 2655 |
+
"kd_loss": 0.4855685773849018,
|
| 2656 |
+
"learning_rate": 3e-06,
|
| 2657 |
+
"loss": 0.5408,
|
| 2658 |
+
"masked_tokens": 110.25,
|
| 2659 |
+
"mean_t": 0.5327763411332853,
|
| 2660 |
+
"step": 1020,
|
| 2661 |
+
"student_masked_tokens": 110.25
|
| 2662 |
+
},
|
| 2663 |
+
{
|
| 2664 |
+
"avg_mask_ratio": 0.46985941788880153,
|
| 2665 |
+
"avg_response_length": 220.05,
|
| 2666 |
+
"avg_student_mask_ratio": 0.46985941788880153,
|
| 2667 |
+
"batch_ainp_frac": 0.0,
|
| 2668 |
+
"batch_inp_frac": 1.0,
|
| 2669 |
+
"batch_inp_oh_frac": 0.0,
|
| 2670 |
+
"batch_inp_par_frac": 0.0,
|
| 2671 |
+
"batch_inp_par_par_frac": 0.0,
|
| 2672 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 2673 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 2674 |
+
"batch_rl_frac": 0.0,
|
| 2675 |
+
"batch_sft_frac": 0.0,
|
| 2676 |
+
"batch_soft_sft_frac": 0.0,
|
| 2677 |
+
"batch_tf_frac": 0.0,
|
| 2678 |
+
"ce_loss": 0.0,
|
| 2679 |
+
"epoch": 2.2005333333333335,
|
| 2680 |
+
"grad_norm": 0.2490234375,
|
| 2681 |
+
"kd_loss": 0.4395623657969203,
|
| 2682 |
+
"learning_rate": 3e-06,
|
| 2683 |
+
"loss": 0.4429,
|
| 2684 |
+
"masked_tokens": 104.9,
|
| 2685 |
+
"mean_t": 0.5033508580760099,
|
| 2686 |
+
"step": 1030,
|
| 2687 |
+
"student_masked_tokens": 104.9
|
| 2688 |
+
},
|
| 2689 |
+
{
|
| 2690 |
+
"avg_mask_ratio": 0.49566771630197765,
|
| 2691 |
+
"avg_response_length": 213.7,
|
| 2692 |
+
"avg_student_mask_ratio": 0.49566771630197765,
|
| 2693 |
+
"batch_ainp_frac": 0.0,
|
| 2694 |
+
"batch_inp_frac": 1.0,
|
| 2695 |
+
"batch_inp_oh_frac": 0.0,
|
| 2696 |
+
"batch_inp_par_frac": 0.0,
|
| 2697 |
+
"batch_inp_par_par_frac": 0.0,
|
| 2698 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 2699 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 2700 |
+
"batch_rl_frac": 0.0,
|
| 2701 |
+
"batch_sft_frac": 0.0,
|
| 2702 |
+
"batch_soft_sft_frac": 0.0,
|
| 2703 |
+
"batch_tf_frac": 0.0,
|
| 2704 |
+
"ce_loss": 0.0,
|
| 2705 |
+
"epoch": 2.2218666666666667,
|
| 2706 |
+
"grad_norm": 0.255859375,
|
| 2707 |
+
"kd_loss": 0.5354410599338735,
|
| 2708 |
+
"learning_rate": 3e-06,
|
| 2709 |
+
"loss": 0.5235,
|
| 2710 |
+
"masked_tokens": 100.35,
|
| 2711 |
+
"mean_t": 0.5349024560535327,
|
| 2712 |
+
"step": 1040,
|
| 2713 |
+
"student_masked_tokens": 100.35
|
| 2714 |
+
},
|
| 2715 |
+
{
|
| 2716 |
+
"avg_mask_ratio": 0.5123252369463444,
|
| 2717 |
+
"avg_response_length": 239.1125,
|
| 2718 |
+
"avg_student_mask_ratio": 0.5123252369463444,
|
| 2719 |
+
"batch_ainp_frac": 0.0,
|
| 2720 |
+
"batch_inp_frac": 1.0,
|
| 2721 |
+
"batch_inp_oh_frac": 0.0,
|
| 2722 |
+
"batch_inp_par_frac": 0.0,
|
| 2723 |
+
"batch_inp_par_par_frac": 0.0,
|
| 2724 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 2725 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 2726 |
+
"batch_rl_frac": 0.0,
|
| 2727 |
+
"batch_sft_frac": 0.0,
|
| 2728 |
+
"batch_soft_sft_frac": 0.0,
|
| 2729 |
+
"batch_tf_frac": 0.0,
|
| 2730 |
+
"ce_loss": 0.0,
|
| 2731 |
+
"epoch": 2.2432,
|
| 2732 |
+
"grad_norm": 0.15625,
|
| 2733 |
+
"kd_loss": 0.5480479164792996,
|
| 2734 |
+
"learning_rate": 3e-06,
|
| 2735 |
+
"loss": 0.4879,
|
| 2736 |
+
"masked_tokens": 123.0,
|
| 2737 |
+
"mean_t": 0.5457118917722255,
|
| 2738 |
+
"step": 1050,
|
| 2739 |
+
"student_masked_tokens": 123.0
|
| 2740 |
+
},
|
| 2741 |
+
{
|
| 2742 |
+
"avg_mask_ratio": 0.46218636581033934,
|
| 2743 |
+
"avg_response_length": 273.7875,
|
| 2744 |
+
"avg_student_mask_ratio": 0.46218636581033934,
|
| 2745 |
+
"batch_ainp_frac": 0.0,
|
| 2746 |
+
"batch_inp_frac": 1.0,
|
| 2747 |
+
"batch_inp_oh_frac": 0.0,
|
| 2748 |
+
"batch_inp_par_frac": 0.0,
|
| 2749 |
+
"batch_inp_par_par_frac": 0.0,
|
| 2750 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 2751 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 2752 |
+
"batch_rl_frac": 0.0,
|
| 2753 |
+
"batch_sft_frac": 0.0,
|
| 2754 |
+
"batch_soft_sft_frac": 0.0,
|
| 2755 |
+
"batch_tf_frac": 0.0,
|
| 2756 |
+
"ce_loss": 0.0,
|
| 2757 |
+
"epoch": 2.2645333333333335,
|
| 2758 |
+
"grad_norm": 0.2314453125,
|
| 2759 |
+
"kd_loss": 0.48405452181636976,
|
| 2760 |
+
"learning_rate": 3e-06,
|
| 2761 |
+
"loss": 0.4551,
|
| 2762 |
+
"masked_tokens": 122.125,
|
| 2763 |
+
"mean_t": 0.48194136443780733,
|
| 2764 |
+
"step": 1060,
|
| 2765 |
+
"student_masked_tokens": 122.125
|
| 2766 |
+
},
|
| 2767 |
+
{
|
| 2768 |
+
"avg_mask_ratio": 0.48546303423354403,
|
| 2769 |
+
"avg_response_length": 260.8125,
|
| 2770 |
+
"avg_student_mask_ratio": 0.48546303423354403,
|
| 2771 |
+
"batch_ainp_frac": 0.0,
|
| 2772 |
+
"batch_inp_frac": 1.0,
|
| 2773 |
+
"batch_inp_oh_frac": 0.0,
|
| 2774 |
+
"batch_inp_par_frac": 0.0,
|
| 2775 |
+
"batch_inp_par_par_frac": 0.0,
|
| 2776 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 2777 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 2778 |
+
"batch_rl_frac": 0.0,
|
| 2779 |
+
"batch_sft_frac": 0.0,
|
| 2780 |
+
"batch_soft_sft_frac": 0.0,
|
| 2781 |
+
"batch_tf_frac": 0.0,
|
| 2782 |
+
"ce_loss": 0.0,
|
| 2783 |
+
"epoch": 2.2858666666666667,
|
| 2784 |
+
"grad_norm": 0.1826171875,
|
| 2785 |
+
"kd_loss": 0.4191161100130927,
|
| 2786 |
+
"learning_rate": 3e-06,
|
| 2787 |
+
"loss": 0.4471,
|
| 2788 |
+
"masked_tokens": 138.0375,
|
| 2789 |
+
"mean_t": 0.5015889146190602,
|
| 2790 |
+
"step": 1070,
|
| 2791 |
+
"student_masked_tokens": 138.0375
|
| 2792 |
+
},
|
| 2793 |
+
{
|
| 2794 |
+
"avg_mask_ratio": 0.4625250873621553,
|
| 2795 |
+
"avg_response_length": 219.8875,
|
| 2796 |
+
"avg_student_mask_ratio": 0.4625250873621553,
|
| 2797 |
+
"batch_ainp_frac": 0.0,
|
| 2798 |
+
"batch_inp_frac": 1.0,
|
| 2799 |
+
"batch_inp_oh_frac": 0.0,
|
| 2800 |
+
"batch_inp_par_frac": 0.0,
|
| 2801 |
+
"batch_inp_par_par_frac": 0.0,
|
| 2802 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 2803 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 2804 |
+
"batch_rl_frac": 0.0,
|
| 2805 |
+
"batch_sft_frac": 0.0,
|
| 2806 |
+
"batch_soft_sft_frac": 0.0,
|
| 2807 |
+
"batch_tf_frac": 0.0,
|
| 2808 |
+
"ce_loss": 0.0,
|
| 2809 |
+
"epoch": 2.3072,
|
| 2810 |
+
"grad_norm": 0.06982421875,
|
| 2811 |
+
"kd_loss": 0.36283817572875704,
|
| 2812 |
+
"learning_rate": 3e-06,
|
| 2813 |
+
"loss": 0.3296,
|
| 2814 |
+
"masked_tokens": 100.5625,
|
| 2815 |
+
"mean_t": 0.4983203248586506,
|
| 2816 |
+
"step": 1080,
|
| 2817 |
+
"student_masked_tokens": 100.5625
|
| 2818 |
+
},
|
| 2819 |
+
{
|
| 2820 |
+
"avg_mask_ratio": 0.44248262273031286,
|
| 2821 |
+
"avg_response_length": 213.55,
|
| 2822 |
+
"avg_student_mask_ratio": 0.44248262273031286,
|
| 2823 |
+
"batch_ainp_frac": 0.0,
|
| 2824 |
+
"batch_inp_frac": 1.0,
|
| 2825 |
+
"batch_inp_oh_frac": 0.0,
|
| 2826 |
+
"batch_inp_par_frac": 0.0,
|
| 2827 |
+
"batch_inp_par_par_frac": 0.0,
|
| 2828 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 2829 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 2830 |
+
"batch_rl_frac": 0.0,
|
| 2831 |
+
"batch_sft_frac": 0.0,
|
| 2832 |
+
"batch_soft_sft_frac": 0.0,
|
| 2833 |
+
"batch_tf_frac": 0.0,
|
| 2834 |
+
"ce_loss": 0.0,
|
| 2835 |
+
"epoch": 2.3285333333333336,
|
| 2836 |
+
"grad_norm": 0.10888671875,
|
| 2837 |
+
"kd_loss": 0.39130586145803364,
|
| 2838 |
+
"learning_rate": 3e-06,
|
| 2839 |
+
"loss": 0.3501,
|
| 2840 |
+
"masked_tokens": 91.9125,
|
| 2841 |
+
"mean_t": 0.47094749807147307,
|
| 2842 |
+
"step": 1090,
|
| 2843 |
+
"student_masked_tokens": 91.9125
|
| 2844 |
+
},
|
| 2845 |
+
{
|
| 2846 |
+
"avg_mask_ratio": 0.5204601250356063,
|
| 2847 |
+
"avg_response_length": 246.1125,
|
| 2848 |
+
"avg_student_mask_ratio": 0.5204601250356063,
|
| 2849 |
+
"batch_ainp_frac": 0.0,
|
| 2850 |
+
"batch_inp_frac": 1.0,
|
| 2851 |
+
"batch_inp_oh_frac": 0.0,
|
| 2852 |
+
"batch_inp_par_frac": 0.0,
|
| 2853 |
+
"batch_inp_par_par_frac": 0.0,
|
| 2854 |
+
"batch_inp_par_reverse_frac": 0.0,
|
| 2855 |
+
"batch_inp_rkd_ace_frac": 0.0,
|
| 2856 |
+
"batch_rl_frac": 0.0,
|
| 2857 |
+
"batch_sft_frac": 0.0,
|
| 2858 |
+
"batch_soft_sft_frac": 0.0,
|
| 2859 |
+
"batch_tf_frac": 0.0,
|
| 2860 |
+
"ce_loss": 0.0,
|
| 2861 |
+
"epoch": 2.3498666666666668,
|
| 2862 |
+
"grad_norm": 0.2177734375,
|
| 2863 |
+
"kd_loss": 0.42874607622297845,
|
| 2864 |
+
"learning_rate": 3e-06,
|
| 2865 |
+
"loss": 0.4962,
|
| 2866 |
+
"masked_tokens": 133.1875,
|
| 2867 |
+
"mean_t": 0.5531192034482956,
|
| 2868 |
+
"step": 1100,
|
| 2869 |
+
"student_masked_tokens": 133.1875
|
| 2870 |
+
}
|
| 2871 |
+
],
|
| 2872 |
+
"logging_steps": 10,
|
| 2873 |
+
"max_steps": 1404,
|
| 2874 |
+
"num_input_tokens_seen": 0,
|
| 2875 |
+
"num_train_epochs": 3,
|
| 2876 |
+
"save_steps": 100,
|
| 2877 |
+
"stateful_callbacks": {
|
| 2878 |
+
"TrainerControl": {
|
| 2879 |
+
"args": {
|
| 2880 |
+
"should_epoch_stop": false,
|
| 2881 |
+
"should_evaluate": false,
|
| 2882 |
+
"should_log": false,
|
| 2883 |
+
"should_save": true,
|
| 2884 |
+
"should_training_stop": false
|
| 2885 |
+
},
|
| 2886 |
+
"attributes": {}
|
| 2887 |
+
}
|
| 2888 |
+
},
|
| 2889 |
+
"total_flos": 0.0,
|
| 2890 |
+
"train_batch_size": 1,
|
| 2891 |
+
"trial_name": null,
|
| 2892 |
+
"trial_params": null
|
| 2893 |
+
}
|
math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-1100/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:99e6e32f3c28324ecee9c50eb13825addacaf6ce31b5b6707966fc1ac725dea3
|
| 3 |
+
size 8312
|
math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-1200/README.md
ADDED
|
@@ -0,0 +1,202 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
base_model: GSAI-ML/LLaDA-8B-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 0.15.1
|
math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-1200/adapter_config.json
ADDED
|
@@ -0,0 +1,39 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"alpha_pattern": {},
|
| 3 |
+
"auto_mapping": null,
|
| 4 |
+
"base_model_name_or_path": "GSAI-ML/LLaDA-8B-Instruct",
|
| 5 |
+
"bias": "none",
|
| 6 |
+
"corda_config": null,
|
| 7 |
+
"eva_config": null,
|
| 8 |
+
"exclude_modules": null,
|
| 9 |
+
"fan_in_fan_out": false,
|
| 10 |
+
"inference_mode": true,
|
| 11 |
+
"init_lora_weights": true,
|
| 12 |
+
"layer_replication": null,
|
| 13 |
+
"layers_pattern": null,
|
| 14 |
+
"layers_to_transform": null,
|
| 15 |
+
"loftq_config": {},
|
| 16 |
+
"lora_alpha": 64,
|
| 17 |
+
"lora_bias": false,
|
| 18 |
+
"lora_dropout": 0.05,
|
| 19 |
+
"megatron_config": null,
|
| 20 |
+
"megatron_core": "megatron.core",
|
| 21 |
+
"modules_to_save": null,
|
| 22 |
+
"peft_type": "LORA",
|
| 23 |
+
"r": 128,
|
| 24 |
+
"rank_pattern": {},
|
| 25 |
+
"revision": null,
|
| 26 |
+
"target_modules": [
|
| 27 |
+
"q_proj",
|
| 28 |
+
"k_proj",
|
| 29 |
+
"o_proj",
|
| 30 |
+
"up_proj",
|
| 31 |
+
"gate_proj",
|
| 32 |
+
"v_proj",
|
| 33 |
+
"down_proj"
|
| 34 |
+
],
|
| 35 |
+
"task_type": "CAUSAL_LM",
|
| 36 |
+
"trainable_token_indices": null,
|
| 37 |
+
"use_dora": false,
|
| 38 |
+
"use_rslora": false
|
| 39 |
+
}
|
math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-1200/adapter_model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:834a225f3813acd598b6e03397b3663cabde61f50d2b90b19fbccfb9739ef7d8
|
| 3 |
+
size 2406624648
|
math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-1200/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:05057c8920c2286042e9d82db71d136106b5b1b675f950cf5b35274453c91b4f
|
| 3 |
+
size 671304442
|
math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-1200/rng_state_0.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1394463a46489e6dce7c0369a296b9effad20c6a87b30dbb892b34b73b5d6365
|
| 3 |
+
size 14512
|
math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-1200/rng_state_1.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4c8dcfaaa26518a9158534ae7671344dd035fcb11f3b40b193e3c3bd47123883
|
| 3 |
+
size 14512
|
math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-1200/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7f8c95a6d9085dfcee1e6620c88ede526366d3a02c5018932b1bc04809c0e0c7
|
| 3 |
+
size 1064
|
math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-1200/trainer_state.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-1200/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:99e6e32f3c28324ecee9c50eb13825addacaf6ce31b5b6707966fc1ac725dea3
|
| 3 |
+
size 8312
|
math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-1300/README.md
ADDED
|
@@ -0,0 +1,202 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
base_model: GSAI-ML/LLaDA-8B-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 0.15.1
|
math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-1300/adapter_config.json
ADDED
|
@@ -0,0 +1,39 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"alpha_pattern": {},
|
| 3 |
+
"auto_mapping": null,
|
| 4 |
+
"base_model_name_or_path": "GSAI-ML/LLaDA-8B-Instruct",
|
| 5 |
+
"bias": "none",
|
| 6 |
+
"corda_config": null,
|
| 7 |
+
"eva_config": null,
|
| 8 |
+
"exclude_modules": null,
|
| 9 |
+
"fan_in_fan_out": false,
|
| 10 |
+
"inference_mode": true,
|
| 11 |
+
"init_lora_weights": true,
|
| 12 |
+
"layer_replication": null,
|
| 13 |
+
"layers_pattern": null,
|
| 14 |
+
"layers_to_transform": null,
|
| 15 |
+
"loftq_config": {},
|
| 16 |
+
"lora_alpha": 64,
|
| 17 |
+
"lora_bias": false,
|
| 18 |
+
"lora_dropout": 0.05,
|
| 19 |
+
"megatron_config": null,
|
| 20 |
+
"megatron_core": "megatron.core",
|
| 21 |
+
"modules_to_save": null,
|
| 22 |
+
"peft_type": "LORA",
|
| 23 |
+
"r": 128,
|
| 24 |
+
"rank_pattern": {},
|
| 25 |
+
"revision": null,
|
| 26 |
+
"target_modules": [
|
| 27 |
+
"q_proj",
|
| 28 |
+
"k_proj",
|
| 29 |
+
"o_proj",
|
| 30 |
+
"up_proj",
|
| 31 |
+
"gate_proj",
|
| 32 |
+
"v_proj",
|
| 33 |
+
"down_proj"
|
| 34 |
+
],
|
| 35 |
+
"task_type": "CAUSAL_LM",
|
| 36 |
+
"trainable_token_indices": null,
|
| 37 |
+
"use_dora": false,
|
| 38 |
+
"use_rslora": false
|
| 39 |
+
}
|
math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-1300/adapter_model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:009f8cfc5daea169c6297616905caf37224d64b635e73b093f4a58186199f9c5
|
| 3 |
+
size 2406624648
|
math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-1300/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4dedb8a15f7507c820788c3e3ff7f5669f5a18f41ac6b2c5357d6a679401c8a8
|
| 3 |
+
size 671304442
|
math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-1300/rng_state_0.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5fed4fb76d2cb63b43a8b71459a39a6ff6dd752f3acf64d0dd3af2ae385d4295
|
| 3 |
+
size 14512
|
math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-1300/rng_state_1.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a37caa81ca3b37bcfae996e60a1ec762e4a75d1e14211368b059806d5e594b2b
|
| 3 |
+
size 14512
|
math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-1300/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:368502bb8b2f2d6f452bda7249e88ca57b330ec2ba407ec248613b4300c99d0d
|
| 3 |
+
size 1064
|
math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-1300/trainer_state.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-1300/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:99e6e32f3c28324ecee9c50eb13825addacaf6ce31b5b6707966fc1ac725dea3
|
| 3 |
+
size 8312
|
math/INP/unmask_tags_leave_last_step_gold1_target1_ce0.0/checkpoint-1400/README.md
ADDED
|
@@ -0,0 +1,202 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
base_model: GSAI-ML/LLaDA-8B-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 0.15.1
|