Upload folder using huggingface_hub
Browse files- .gitattributes +4 -0
- checkpoint-200/adapter_config.json +42 -0
- checkpoint-200/special_tokens_map.json +24 -0
- checkpoint-200/tokenizer.json +3 -0
- checkpoint-200/tokenizer_config.json +0 -0
- checkpoint-200/trainer_state.json +278 -0
- checkpoint-250/adapter_config.json +42 -0
- checkpoint-250/special_tokens_map.json +24 -0
- checkpoint-250/tokenizer.json +3 -0
- checkpoint-250/tokenizer_config.json +0 -0
- checkpoint-250/trainer_state.json +339 -0
- checkpoint-500/adapter_config.json +42 -0
- checkpoint-500/special_tokens_map.json +24 -0
- checkpoint-500/tokenizer.json +3 -0
- checkpoint-500/tokenizer_config.json +0 -0
- checkpoint-500/trainer_state.json +644 -0
- special_tokens_map.json +24 -0
- tokenizer.json +3 -0
- tokenizer_config.json +0 -0
.gitattributes
CHANGED
|
@@ -33,3 +33,7 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
|
|
| 33 |
*.zip filter=lfs diff=lfs merge=lfs -text
|
| 34 |
*.zst filter=lfs diff=lfs merge=lfs -text
|
| 35 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 33 |
*.zip filter=lfs diff=lfs merge=lfs -text
|
| 34 |
*.zst filter=lfs diff=lfs merge=lfs -text
|
| 35 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
| 36 |
+
checkpoint-200/tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
| 37 |
+
checkpoint-250/tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
| 38 |
+
checkpoint-500/tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
| 39 |
+
tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
checkpoint-200/adapter_config.json
ADDED
|
@@ -0,0 +1,42 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"alpha_pattern": {},
|
| 3 |
+
"auto_mapping": null,
|
| 4 |
+
"base_model_name_or_path": "/home/support/llm/Ministral-8B-Instruct-2410/",
|
| 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": 16,
|
| 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 |
+
"qalora_group_size": 16,
|
| 24 |
+
"r": 64,
|
| 25 |
+
"rank_pattern": {},
|
| 26 |
+
"revision": null,
|
| 27 |
+
"target_modules": [
|
| 28 |
+
"up_proj",
|
| 29 |
+
"k_proj",
|
| 30 |
+
"down_proj",
|
| 31 |
+
"q_proj",
|
| 32 |
+
"gate_proj",
|
| 33 |
+
"o_proj",
|
| 34 |
+
"v_proj"
|
| 35 |
+
],
|
| 36 |
+
"target_parameters": null,
|
| 37 |
+
"task_type": "CAUSAL_LM",
|
| 38 |
+
"trainable_token_indices": null,
|
| 39 |
+
"use_dora": false,
|
| 40 |
+
"use_qalora": false,
|
| 41 |
+
"use_rslora": false
|
| 42 |
+
}
|
checkpoint-200/special_tokens_map.json
ADDED
|
@@ -0,0 +1,24 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"content": "<s>",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": false,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"eos_token": {
|
| 10 |
+
"content": "</s>",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": false,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"pad_token": "</s>",
|
| 17 |
+
"unk_token": {
|
| 18 |
+
"content": "<unk>",
|
| 19 |
+
"lstrip": false,
|
| 20 |
+
"normalized": false,
|
| 21 |
+
"rstrip": false,
|
| 22 |
+
"single_word": false
|
| 23 |
+
}
|
| 24 |
+
}
|
checkpoint-200/tokenizer.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d7edbeaf20dd7f571b5dd1c54d9ace4f9b6299127cc7ba2afb14a6d51a4a79a4
|
| 3 |
+
size 17078136
|
checkpoint-200/tokenizer_config.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
checkpoint-200/trainer_state.json
ADDED
|
@@ -0,0 +1,278 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 200,
|
| 3 |
+
"best_metric": 0.6800833940505981,
|
| 4 |
+
"best_model_checkpoint": "./python-code-completion-ministral-8B/checkpoint-200",
|
| 5 |
+
"epoch": 0.9411764705882353,
|
| 6 |
+
"eval_steps": 50,
|
| 7 |
+
"global_step": 200,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"entropy": 1.4469833582639695,
|
| 14 |
+
"epoch": 0.047058823529411764,
|
| 15 |
+
"grad_norm": 0.49369218945503235,
|
| 16 |
+
"learning_rate": 9.999578688879085e-06,
|
| 17 |
+
"loss": 1.7076,
|
| 18 |
+
"mean_token_accuracy": 0.6342830918729305,
|
| 19 |
+
"num_tokens": 25138.0,
|
| 20 |
+
"step": 10
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"entropy": 1.430517853796482,
|
| 24 |
+
"epoch": 0.09411764705882353,
|
| 25 |
+
"grad_norm": 0.51407390832901,
|
| 26 |
+
"learning_rate": 9.949107209404664e-06,
|
| 27 |
+
"loss": 1.5636,
|
| 28 |
+
"mean_token_accuracy": 0.6443205893039703,
|
| 29 |
+
"num_tokens": 47578.0,
|
| 30 |
+
"step": 20
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"entropy": 1.4371182411909103,
|
| 34 |
+
"epoch": 0.1411764705882353,
|
| 35 |
+
"grad_norm": 0.6639813184738159,
|
| 36 |
+
"learning_rate": 9.815347070757234e-06,
|
| 37 |
+
"loss": 1.4756,
|
| 38 |
+
"mean_token_accuracy": 0.6693132996559144,
|
| 39 |
+
"num_tokens": 71393.0,
|
| 40 |
+
"step": 30
|
| 41 |
+
},
|
| 42 |
+
{
|
| 43 |
+
"entropy": 1.3472976341843605,
|
| 44 |
+
"epoch": 0.18823529411764706,
|
| 45 |
+
"grad_norm": 0.3675422668457031,
|
| 46 |
+
"learning_rate": 9.60054932593026e-06,
|
| 47 |
+
"loss": 1.2256,
|
| 48 |
+
"mean_token_accuracy": 0.6993286892771721,
|
| 49 |
+
"num_tokens": 96359.0,
|
| 50 |
+
"step": 40
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"entropy": 1.1503819346427917,
|
| 54 |
+
"epoch": 0.23529411764705882,
|
| 55 |
+
"grad_norm": 0.4163525104522705,
|
| 56 |
+
"learning_rate": 9.30832881198487e-06,
|
| 57 |
+
"loss": 1.0136,
|
| 58 |
+
"mean_token_accuracy": 0.7729795873165131,
|
| 59 |
+
"num_tokens": 123155.0,
|
| 60 |
+
"step": 50
|
| 61 |
+
},
|
| 62 |
+
{
|
| 63 |
+
"epoch": 0.23529411764705882,
|
| 64 |
+
"eval_entropy": 1.0332347961266835,
|
| 65 |
+
"eval_loss": 0.9837512373924255,
|
| 66 |
+
"eval_mean_token_accuracy": 0.8083523658911387,
|
| 67 |
+
"eval_num_tokens": 123155.0,
|
| 68 |
+
"eval_runtime": 25.0181,
|
| 69 |
+
"eval_samples_per_second": 5.996,
|
| 70 |
+
"eval_steps_per_second": 5.996,
|
| 71 |
+
"step": 50
|
| 72 |
+
},
|
| 73 |
+
{
|
| 74 |
+
"entropy": 0.8932387962937355,
|
| 75 |
+
"epoch": 0.2823529411764706,
|
| 76 |
+
"grad_norm": 0.2984226942062378,
|
| 77 |
+
"learning_rate": 8.943603315859101e-06,
|
| 78 |
+
"loss": 0.8519,
|
| 79 |
+
"mean_token_accuracy": 0.8314538225531578,
|
| 80 |
+
"num_tokens": 149237.0,
|
| 81 |
+
"step": 60
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"entropy": 0.7480494946241378,
|
| 85 |
+
"epoch": 0.32941176470588235,
|
| 86 |
+
"grad_norm": 0.409252405166626,
|
| 87 |
+
"learning_rate": 8.512510812798426e-06,
|
| 88 |
+
"loss": 0.7625,
|
| 89 |
+
"mean_token_accuracy": 0.853736138343811,
|
| 90 |
+
"num_tokens": 173954.0,
|
| 91 |
+
"step": 70
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"entropy": 0.733599005267024,
|
| 95 |
+
"epoch": 0.3764705882352941,
|
| 96 |
+
"grad_norm": 0.3177991211414337,
|
| 97 |
+
"learning_rate": 8.022306170204233e-06,
|
| 98 |
+
"loss": 0.7384,
|
| 99 |
+
"mean_token_accuracy": 0.8508381456136703,
|
| 100 |
+
"num_tokens": 202672.0,
|
| 101 |
+
"step": 80
|
| 102 |
+
},
|
| 103 |
+
{
|
| 104 |
+
"entropy": 0.6522206105291843,
|
| 105 |
+
"epoch": 0.4235294117647059,
|
| 106 |
+
"grad_norm": 0.5717107057571411,
|
| 107 |
+
"learning_rate": 7.481239055273959e-06,
|
| 108 |
+
"loss": 0.6701,
|
| 109 |
+
"mean_token_accuracy": 0.8760824427008629,
|
| 110 |
+
"num_tokens": 224695.0,
|
| 111 |
+
"step": 90
|
| 112 |
+
},
|
| 113 |
+
{
|
| 114 |
+
"entropy": 0.7285707335919142,
|
| 115 |
+
"epoch": 0.47058823529411764,
|
| 116 |
+
"grad_norm": 0.21871985495090485,
|
| 117 |
+
"learning_rate": 6.898415101128571e-06,
|
| 118 |
+
"loss": 0.7319,
|
| 119 |
+
"mean_token_accuracy": 0.8523965805768967,
|
| 120 |
+
"num_tokens": 251370.0,
|
| 121 |
+
"step": 100
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.47058823529411764,
|
| 125 |
+
"eval_entropy": 0.73099374294281,
|
| 126 |
+
"eval_loss": 0.7096962332725525,
|
| 127 |
+
"eval_mean_token_accuracy": 0.8550362022717793,
|
| 128 |
+
"eval_num_tokens": 251370.0,
|
| 129 |
+
"eval_runtime": 25.1112,
|
| 130 |
+
"eval_samples_per_second": 5.973,
|
| 131 |
+
"eval_steps_per_second": 5.973,
|
| 132 |
+
"step": 100
|
| 133 |
+
},
|
| 134 |
+
{
|
| 135 |
+
"entropy": 0.6658912554383278,
|
| 136 |
+
"epoch": 0.5176470588235295,
|
| 137 |
+
"grad_norm": 0.2736637592315674,
|
| 138 |
+
"learning_rate": 6.283642667866317e-06,
|
| 139 |
+
"loss": 0.6646,
|
| 140 |
+
"mean_token_accuracy": 0.8663464844226837,
|
| 141 |
+
"num_tokens": 276477.0,
|
| 142 |
+
"step": 110
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"entropy": 0.6516228564083576,
|
| 146 |
+
"epoch": 0.5647058823529412,
|
| 147 |
+
"grad_norm": 0.2847215533256531,
|
| 148 |
+
"learning_rate": 5.647267777405178e-06,
|
| 149 |
+
"loss": 0.6503,
|
| 150 |
+
"mean_token_accuracy": 0.8681117251515389,
|
| 151 |
+
"num_tokens": 302513.0,
|
| 152 |
+
"step": 120
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"entropy": 0.7009131267666817,
|
| 156 |
+
"epoch": 0.611764705882353,
|
| 157 |
+
"grad_norm": 0.22739574313163757,
|
| 158 |
+
"learning_rate": 5e-06,
|
| 159 |
+
"loss": 0.6881,
|
| 160 |
+
"mean_token_accuracy": 0.8624396547675133,
|
| 161 |
+
"num_tokens": 329679.0,
|
| 162 |
+
"step": 130
|
| 163 |
+
},
|
| 164 |
+
{
|
| 165 |
+
"entropy": 0.6468490872532129,
|
| 166 |
+
"epoch": 0.6588235294117647,
|
| 167 |
+
"grad_norm": 0.3506576716899872,
|
| 168 |
+
"learning_rate": 4.352732222594823e-06,
|
| 169 |
+
"loss": 0.6462,
|
| 170 |
+
"mean_token_accuracy": 0.871060524880886,
|
| 171 |
+
"num_tokens": 353837.0,
|
| 172 |
+
"step": 140
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"entropy": 0.6587031967937946,
|
| 176 |
+
"epoch": 0.7058823529411765,
|
| 177 |
+
"grad_norm": 0.2826354205608368,
|
| 178 |
+
"learning_rate": 3.7163573321336867e-06,
|
| 179 |
+
"loss": 0.6779,
|
| 180 |
+
"mean_token_accuracy": 0.8682561233639717,
|
| 181 |
+
"num_tokens": 376786.0,
|
| 182 |
+
"step": 150
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 0.7058823529411765,
|
| 186 |
+
"eval_entropy": 0.7037040118376414,
|
| 187 |
+
"eval_loss": 0.6883785724639893,
|
| 188 |
+
"eval_mean_token_accuracy": 0.8590577419598897,
|
| 189 |
+
"eval_num_tokens": 376786.0,
|
| 190 |
+
"eval_runtime": 25.1113,
|
| 191 |
+
"eval_samples_per_second": 5.973,
|
| 192 |
+
"eval_steps_per_second": 5.973,
|
| 193 |
+
"step": 150
|
| 194 |
+
},
|
| 195 |
+
{
|
| 196 |
+
"entropy": 0.6935035854578018,
|
| 197 |
+
"epoch": 0.7529411764705882,
|
| 198 |
+
"grad_norm": 0.32049205899238586,
|
| 199 |
+
"learning_rate": 3.101584898871431e-06,
|
| 200 |
+
"loss": 0.7223,
|
| 201 |
+
"mean_token_accuracy": 0.8527676403522492,
|
| 202 |
+
"num_tokens": 404164.0,
|
| 203 |
+
"step": 160
|
| 204 |
+
},
|
| 205 |
+
{
|
| 206 |
+
"entropy": 0.6355765651911497,
|
| 207 |
+
"epoch": 0.8,
|
| 208 |
+
"grad_norm": 0.27511146664619446,
|
| 209 |
+
"learning_rate": 2.518760944726042e-06,
|
| 210 |
+
"loss": 0.641,
|
| 211 |
+
"mean_token_accuracy": 0.8753799632191658,
|
| 212 |
+
"num_tokens": 427356.0,
|
| 213 |
+
"step": 170
|
| 214 |
+
},
|
| 215 |
+
{
|
| 216 |
+
"entropy": 0.6331917632371187,
|
| 217 |
+
"epoch": 0.8470588235294118,
|
| 218 |
+
"grad_norm": 0.289739191532135,
|
| 219 |
+
"learning_rate": 1.977693829795769e-06,
|
| 220 |
+
"loss": 0.646,
|
| 221 |
+
"mean_token_accuracy": 0.8754393547773361,
|
| 222 |
+
"num_tokens": 450321.0,
|
| 223 |
+
"step": 180
|
| 224 |
+
},
|
| 225 |
+
{
|
| 226 |
+
"entropy": 0.7034022852778434,
|
| 227 |
+
"epoch": 0.8941176470588236,
|
| 228 |
+
"grad_norm": 0.32424843311309814,
|
| 229 |
+
"learning_rate": 1.4874891872015735e-06,
|
| 230 |
+
"loss": 0.7084,
|
| 231 |
+
"mean_token_accuracy": 0.857265616953373,
|
| 232 |
+
"num_tokens": 478364.0,
|
| 233 |
+
"step": 190
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"entropy": 0.6966812558472156,
|
| 237 |
+
"epoch": 0.9411764705882353,
|
| 238 |
+
"grad_norm": 0.27991771697998047,
|
| 239 |
+
"learning_rate": 1.0563966841408995e-06,
|
| 240 |
+
"loss": 0.7108,
|
| 241 |
+
"mean_token_accuracy": 0.857306943833828,
|
| 242 |
+
"num_tokens": 505339.0,
|
| 243 |
+
"step": 200
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"epoch": 0.9411764705882353,
|
| 247 |
+
"eval_entropy": 0.6970382163921992,
|
| 248 |
+
"eval_loss": 0.6800833940505981,
|
| 249 |
+
"eval_mean_token_accuracy": 0.8598487377166748,
|
| 250 |
+
"eval_num_tokens": 505339.0,
|
| 251 |
+
"eval_runtime": 25.145,
|
| 252 |
+
"eval_samples_per_second": 5.965,
|
| 253 |
+
"eval_steps_per_second": 5.965,
|
| 254 |
+
"step": 200
|
| 255 |
+
}
|
| 256 |
+
],
|
| 257 |
+
"logging_steps": 10,
|
| 258 |
+
"max_steps": 250,
|
| 259 |
+
"num_input_tokens_seen": 0,
|
| 260 |
+
"num_train_epochs": 2,
|
| 261 |
+
"save_steps": 100,
|
| 262 |
+
"stateful_callbacks": {
|
| 263 |
+
"TrainerControl": {
|
| 264 |
+
"args": {
|
| 265 |
+
"should_epoch_stop": false,
|
| 266 |
+
"should_evaluate": false,
|
| 267 |
+
"should_log": false,
|
| 268 |
+
"should_save": true,
|
| 269 |
+
"should_training_stop": false
|
| 270 |
+
},
|
| 271 |
+
"attributes": {}
|
| 272 |
+
}
|
| 273 |
+
},
|
| 274 |
+
"total_flos": 2.321787690779443e+16,
|
| 275 |
+
"train_batch_size": 1,
|
| 276 |
+
"trial_name": null,
|
| 277 |
+
"trial_params": null
|
| 278 |
+
}
|
checkpoint-250/adapter_config.json
ADDED
|
@@ -0,0 +1,42 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"alpha_pattern": {},
|
| 3 |
+
"auto_mapping": null,
|
| 4 |
+
"base_model_name_or_path": "/home/support/llm/Ministral-8B-Instruct-2410/",
|
| 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": 16,
|
| 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 |
+
"qalora_group_size": 16,
|
| 24 |
+
"r": 64,
|
| 25 |
+
"rank_pattern": {},
|
| 26 |
+
"revision": null,
|
| 27 |
+
"target_modules": [
|
| 28 |
+
"up_proj",
|
| 29 |
+
"k_proj",
|
| 30 |
+
"down_proj",
|
| 31 |
+
"q_proj",
|
| 32 |
+
"gate_proj",
|
| 33 |
+
"o_proj",
|
| 34 |
+
"v_proj"
|
| 35 |
+
],
|
| 36 |
+
"target_parameters": null,
|
| 37 |
+
"task_type": "CAUSAL_LM",
|
| 38 |
+
"trainable_token_indices": null,
|
| 39 |
+
"use_dora": false,
|
| 40 |
+
"use_qalora": false,
|
| 41 |
+
"use_rslora": false
|
| 42 |
+
}
|
checkpoint-250/special_tokens_map.json
ADDED
|
@@ -0,0 +1,24 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"content": "<s>",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": false,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"eos_token": {
|
| 10 |
+
"content": "</s>",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": false,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"pad_token": "</s>",
|
| 17 |
+
"unk_token": {
|
| 18 |
+
"content": "<unk>",
|
| 19 |
+
"lstrip": false,
|
| 20 |
+
"normalized": false,
|
| 21 |
+
"rstrip": false,
|
| 22 |
+
"single_word": false
|
| 23 |
+
}
|
| 24 |
+
}
|
checkpoint-250/tokenizer.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d7edbeaf20dd7f571b5dd1c54d9ace4f9b6299127cc7ba2afb14a6d51a4a79a4
|
| 3 |
+
size 17078136
|
checkpoint-250/tokenizer_config.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
checkpoint-250/trainer_state.json
ADDED
|
@@ -0,0 +1,339 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 250,
|
| 3 |
+
"best_metric": 0.678814709186554,
|
| 4 |
+
"best_model_checkpoint": "./python-code-completion-ministral-8B/checkpoint-250",
|
| 5 |
+
"epoch": 1.1741176470588235,
|
| 6 |
+
"eval_steps": 50,
|
| 7 |
+
"global_step": 250,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"entropy": 1.4469833582639695,
|
| 14 |
+
"epoch": 0.047058823529411764,
|
| 15 |
+
"grad_norm": 0.49369218945503235,
|
| 16 |
+
"learning_rate": 9.999578688879085e-06,
|
| 17 |
+
"loss": 1.7076,
|
| 18 |
+
"mean_token_accuracy": 0.6342830918729305,
|
| 19 |
+
"num_tokens": 25138.0,
|
| 20 |
+
"step": 10
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"entropy": 1.430517853796482,
|
| 24 |
+
"epoch": 0.09411764705882353,
|
| 25 |
+
"grad_norm": 0.51407390832901,
|
| 26 |
+
"learning_rate": 9.949107209404664e-06,
|
| 27 |
+
"loss": 1.5636,
|
| 28 |
+
"mean_token_accuracy": 0.6443205893039703,
|
| 29 |
+
"num_tokens": 47578.0,
|
| 30 |
+
"step": 20
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"entropy": 1.4371182411909103,
|
| 34 |
+
"epoch": 0.1411764705882353,
|
| 35 |
+
"grad_norm": 0.6639813184738159,
|
| 36 |
+
"learning_rate": 9.815347070757234e-06,
|
| 37 |
+
"loss": 1.4756,
|
| 38 |
+
"mean_token_accuracy": 0.6693132996559144,
|
| 39 |
+
"num_tokens": 71393.0,
|
| 40 |
+
"step": 30
|
| 41 |
+
},
|
| 42 |
+
{
|
| 43 |
+
"entropy": 1.3472976341843605,
|
| 44 |
+
"epoch": 0.18823529411764706,
|
| 45 |
+
"grad_norm": 0.3675422668457031,
|
| 46 |
+
"learning_rate": 9.60054932593026e-06,
|
| 47 |
+
"loss": 1.2256,
|
| 48 |
+
"mean_token_accuracy": 0.6993286892771721,
|
| 49 |
+
"num_tokens": 96359.0,
|
| 50 |
+
"step": 40
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"entropy": 1.1503819346427917,
|
| 54 |
+
"epoch": 0.23529411764705882,
|
| 55 |
+
"grad_norm": 0.4163525104522705,
|
| 56 |
+
"learning_rate": 9.30832881198487e-06,
|
| 57 |
+
"loss": 1.0136,
|
| 58 |
+
"mean_token_accuracy": 0.7729795873165131,
|
| 59 |
+
"num_tokens": 123155.0,
|
| 60 |
+
"step": 50
|
| 61 |
+
},
|
| 62 |
+
{
|
| 63 |
+
"epoch": 0.23529411764705882,
|
| 64 |
+
"eval_entropy": 1.0332347961266835,
|
| 65 |
+
"eval_loss": 0.9837512373924255,
|
| 66 |
+
"eval_mean_token_accuracy": 0.8083523658911387,
|
| 67 |
+
"eval_num_tokens": 123155.0,
|
| 68 |
+
"eval_runtime": 25.0181,
|
| 69 |
+
"eval_samples_per_second": 5.996,
|
| 70 |
+
"eval_steps_per_second": 5.996,
|
| 71 |
+
"step": 50
|
| 72 |
+
},
|
| 73 |
+
{
|
| 74 |
+
"entropy": 0.8932387962937355,
|
| 75 |
+
"epoch": 0.2823529411764706,
|
| 76 |
+
"grad_norm": 0.2984226942062378,
|
| 77 |
+
"learning_rate": 8.943603315859101e-06,
|
| 78 |
+
"loss": 0.8519,
|
| 79 |
+
"mean_token_accuracy": 0.8314538225531578,
|
| 80 |
+
"num_tokens": 149237.0,
|
| 81 |
+
"step": 60
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"entropy": 0.7480494946241378,
|
| 85 |
+
"epoch": 0.32941176470588235,
|
| 86 |
+
"grad_norm": 0.409252405166626,
|
| 87 |
+
"learning_rate": 8.512510812798426e-06,
|
| 88 |
+
"loss": 0.7625,
|
| 89 |
+
"mean_token_accuracy": 0.853736138343811,
|
| 90 |
+
"num_tokens": 173954.0,
|
| 91 |
+
"step": 70
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"entropy": 0.733599005267024,
|
| 95 |
+
"epoch": 0.3764705882352941,
|
| 96 |
+
"grad_norm": 0.3177991211414337,
|
| 97 |
+
"learning_rate": 8.022306170204233e-06,
|
| 98 |
+
"loss": 0.7384,
|
| 99 |
+
"mean_token_accuracy": 0.8508381456136703,
|
| 100 |
+
"num_tokens": 202672.0,
|
| 101 |
+
"step": 80
|
| 102 |
+
},
|
| 103 |
+
{
|
| 104 |
+
"entropy": 0.6522206105291843,
|
| 105 |
+
"epoch": 0.4235294117647059,
|
| 106 |
+
"grad_norm": 0.5717107057571411,
|
| 107 |
+
"learning_rate": 7.481239055273959e-06,
|
| 108 |
+
"loss": 0.6701,
|
| 109 |
+
"mean_token_accuracy": 0.8760824427008629,
|
| 110 |
+
"num_tokens": 224695.0,
|
| 111 |
+
"step": 90
|
| 112 |
+
},
|
| 113 |
+
{
|
| 114 |
+
"entropy": 0.7285707335919142,
|
| 115 |
+
"epoch": 0.47058823529411764,
|
| 116 |
+
"grad_norm": 0.21871985495090485,
|
| 117 |
+
"learning_rate": 6.898415101128571e-06,
|
| 118 |
+
"loss": 0.7319,
|
| 119 |
+
"mean_token_accuracy": 0.8523965805768967,
|
| 120 |
+
"num_tokens": 251370.0,
|
| 121 |
+
"step": 100
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.47058823529411764,
|
| 125 |
+
"eval_entropy": 0.73099374294281,
|
| 126 |
+
"eval_loss": 0.7096962332725525,
|
| 127 |
+
"eval_mean_token_accuracy": 0.8550362022717793,
|
| 128 |
+
"eval_num_tokens": 251370.0,
|
| 129 |
+
"eval_runtime": 25.1112,
|
| 130 |
+
"eval_samples_per_second": 5.973,
|
| 131 |
+
"eval_steps_per_second": 5.973,
|
| 132 |
+
"step": 100
|
| 133 |
+
},
|
| 134 |
+
{
|
| 135 |
+
"entropy": 0.6658912554383278,
|
| 136 |
+
"epoch": 0.5176470588235295,
|
| 137 |
+
"grad_norm": 0.2736637592315674,
|
| 138 |
+
"learning_rate": 6.283642667866317e-06,
|
| 139 |
+
"loss": 0.6646,
|
| 140 |
+
"mean_token_accuracy": 0.8663464844226837,
|
| 141 |
+
"num_tokens": 276477.0,
|
| 142 |
+
"step": 110
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"entropy": 0.6516228564083576,
|
| 146 |
+
"epoch": 0.5647058823529412,
|
| 147 |
+
"grad_norm": 0.2847215533256531,
|
| 148 |
+
"learning_rate": 5.647267777405178e-06,
|
| 149 |
+
"loss": 0.6503,
|
| 150 |
+
"mean_token_accuracy": 0.8681117251515389,
|
| 151 |
+
"num_tokens": 302513.0,
|
| 152 |
+
"step": 120
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"entropy": 0.7009131267666817,
|
| 156 |
+
"epoch": 0.611764705882353,
|
| 157 |
+
"grad_norm": 0.22739574313163757,
|
| 158 |
+
"learning_rate": 5e-06,
|
| 159 |
+
"loss": 0.6881,
|
| 160 |
+
"mean_token_accuracy": 0.8624396547675133,
|
| 161 |
+
"num_tokens": 329679.0,
|
| 162 |
+
"step": 130
|
| 163 |
+
},
|
| 164 |
+
{
|
| 165 |
+
"entropy": 0.6468490872532129,
|
| 166 |
+
"epoch": 0.6588235294117647,
|
| 167 |
+
"grad_norm": 0.3506576716899872,
|
| 168 |
+
"learning_rate": 4.352732222594823e-06,
|
| 169 |
+
"loss": 0.6462,
|
| 170 |
+
"mean_token_accuracy": 0.871060524880886,
|
| 171 |
+
"num_tokens": 353837.0,
|
| 172 |
+
"step": 140
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"entropy": 0.6587031967937946,
|
| 176 |
+
"epoch": 0.7058823529411765,
|
| 177 |
+
"grad_norm": 0.2826354205608368,
|
| 178 |
+
"learning_rate": 3.7163573321336867e-06,
|
| 179 |
+
"loss": 0.6779,
|
| 180 |
+
"mean_token_accuracy": 0.8682561233639717,
|
| 181 |
+
"num_tokens": 376786.0,
|
| 182 |
+
"step": 150
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 0.7058823529411765,
|
| 186 |
+
"eval_entropy": 0.7037040118376414,
|
| 187 |
+
"eval_loss": 0.6883785724639893,
|
| 188 |
+
"eval_mean_token_accuracy": 0.8590577419598897,
|
| 189 |
+
"eval_num_tokens": 376786.0,
|
| 190 |
+
"eval_runtime": 25.1113,
|
| 191 |
+
"eval_samples_per_second": 5.973,
|
| 192 |
+
"eval_steps_per_second": 5.973,
|
| 193 |
+
"step": 150
|
| 194 |
+
},
|
| 195 |
+
{
|
| 196 |
+
"entropy": 0.6935035854578018,
|
| 197 |
+
"epoch": 0.7529411764705882,
|
| 198 |
+
"grad_norm": 0.32049205899238586,
|
| 199 |
+
"learning_rate": 3.101584898871431e-06,
|
| 200 |
+
"loss": 0.7223,
|
| 201 |
+
"mean_token_accuracy": 0.8527676403522492,
|
| 202 |
+
"num_tokens": 404164.0,
|
| 203 |
+
"step": 160
|
| 204 |
+
},
|
| 205 |
+
{
|
| 206 |
+
"entropy": 0.6355765651911497,
|
| 207 |
+
"epoch": 0.8,
|
| 208 |
+
"grad_norm": 0.27511146664619446,
|
| 209 |
+
"learning_rate": 2.518760944726042e-06,
|
| 210 |
+
"loss": 0.641,
|
| 211 |
+
"mean_token_accuracy": 0.8753799632191658,
|
| 212 |
+
"num_tokens": 427356.0,
|
| 213 |
+
"step": 170
|
| 214 |
+
},
|
| 215 |
+
{
|
| 216 |
+
"entropy": 0.6331917632371187,
|
| 217 |
+
"epoch": 0.8470588235294118,
|
| 218 |
+
"grad_norm": 0.289739191532135,
|
| 219 |
+
"learning_rate": 1.977693829795769e-06,
|
| 220 |
+
"loss": 0.646,
|
| 221 |
+
"mean_token_accuracy": 0.8754393547773361,
|
| 222 |
+
"num_tokens": 450321.0,
|
| 223 |
+
"step": 180
|
| 224 |
+
},
|
| 225 |
+
{
|
| 226 |
+
"entropy": 0.7034022852778434,
|
| 227 |
+
"epoch": 0.8941176470588236,
|
| 228 |
+
"grad_norm": 0.32424843311309814,
|
| 229 |
+
"learning_rate": 1.4874891872015735e-06,
|
| 230 |
+
"loss": 0.7084,
|
| 231 |
+
"mean_token_accuracy": 0.857265616953373,
|
| 232 |
+
"num_tokens": 478364.0,
|
| 233 |
+
"step": 190
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"entropy": 0.6966812558472156,
|
| 237 |
+
"epoch": 0.9411764705882353,
|
| 238 |
+
"grad_norm": 0.27991771697998047,
|
| 239 |
+
"learning_rate": 1.0563966841408995e-06,
|
| 240 |
+
"loss": 0.7108,
|
| 241 |
+
"mean_token_accuracy": 0.857306943833828,
|
| 242 |
+
"num_tokens": 505339.0,
|
| 243 |
+
"step": 200
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"epoch": 0.9411764705882353,
|
| 247 |
+
"eval_entropy": 0.6970382163921992,
|
| 248 |
+
"eval_loss": 0.6800833940505981,
|
| 249 |
+
"eval_mean_token_accuracy": 0.8598487377166748,
|
| 250 |
+
"eval_num_tokens": 505339.0,
|
| 251 |
+
"eval_runtime": 25.145,
|
| 252 |
+
"eval_samples_per_second": 5.965,
|
| 253 |
+
"eval_steps_per_second": 5.965,
|
| 254 |
+
"step": 200
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"entropy": 0.7158303387463093,
|
| 258 |
+
"epoch": 0.9882352941176471,
|
| 259 |
+
"grad_norm": 0.41559284925460815,
|
| 260 |
+
"learning_rate": 6.916711880151305e-07,
|
| 261 |
+
"loss": 0.716,
|
| 262 |
+
"mean_token_accuracy": 0.8561157166957856,
|
| 263 |
+
"num_tokens": 533550.0,
|
| 264 |
+
"step": 210
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"entropy": 0.6228710942362484,
|
| 268 |
+
"epoch": 1.0329411764705883,
|
| 269 |
+
"grad_norm": 0.5400373935699463,
|
| 270 |
+
"learning_rate": 3.9945067406974067e-07,
|
| 271 |
+
"loss": 0.6264,
|
| 272 |
+
"mean_token_accuracy": 0.8763874330018696,
|
| 273 |
+
"num_tokens": 552959.0,
|
| 274 |
+
"step": 220
|
| 275 |
+
},
|
| 276 |
+
{
|
| 277 |
+
"entropy": 0.6531048402190208,
|
| 278 |
+
"epoch": 1.08,
|
| 279 |
+
"grad_norm": 0.24900081753730774,
|
| 280 |
+
"learning_rate": 1.8465292924276844e-07,
|
| 281 |
+
"loss": 0.653,
|
| 282 |
+
"mean_token_accuracy": 0.8739597767591476,
|
| 283 |
+
"num_tokens": 575104.0,
|
| 284 |
+
"step": 230
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"entropy": 0.6037626899778843,
|
| 288 |
+
"epoch": 1.1270588235294117,
|
| 289 |
+
"grad_norm": 0.3135730028152466,
|
| 290 |
+
"learning_rate": 5.089279059533658e-08,
|
| 291 |
+
"loss": 0.6126,
|
| 292 |
+
"mean_token_accuracy": 0.8747931569814682,
|
| 293 |
+
"num_tokens": 603807.0,
|
| 294 |
+
"step": 240
|
| 295 |
+
},
|
| 296 |
+
{
|
| 297 |
+
"entropy": 0.665476355701685,
|
| 298 |
+
"epoch": 1.1741176470588235,
|
| 299 |
+
"grad_norm": 0.26424089074134827,
|
| 300 |
+
"learning_rate": 4.213111209155907e-10,
|
| 301 |
+
"loss": 0.6645,
|
| 302 |
+
"mean_token_accuracy": 0.8687899306416511,
|
| 303 |
+
"num_tokens": 628367.0,
|
| 304 |
+
"step": 250
|
| 305 |
+
},
|
| 306 |
+
{
|
| 307 |
+
"epoch": 1.1741176470588235,
|
| 308 |
+
"eval_entropy": 0.6952528850237528,
|
| 309 |
+
"eval_loss": 0.678814709186554,
|
| 310 |
+
"eval_mean_token_accuracy": 0.8600315936406453,
|
| 311 |
+
"eval_num_tokens": 628367.0,
|
| 312 |
+
"eval_runtime": 25.1455,
|
| 313 |
+
"eval_samples_per_second": 5.965,
|
| 314 |
+
"eval_steps_per_second": 5.965,
|
| 315 |
+
"step": 250
|
| 316 |
+
}
|
| 317 |
+
],
|
| 318 |
+
"logging_steps": 10,
|
| 319 |
+
"max_steps": 250,
|
| 320 |
+
"num_input_tokens_seen": 0,
|
| 321 |
+
"num_train_epochs": 2,
|
| 322 |
+
"save_steps": 100,
|
| 323 |
+
"stateful_callbacks": {
|
| 324 |
+
"TrainerControl": {
|
| 325 |
+
"args": {
|
| 326 |
+
"should_epoch_stop": false,
|
| 327 |
+
"should_evaluate": false,
|
| 328 |
+
"should_log": false,
|
| 329 |
+
"should_save": true,
|
| 330 |
+
"should_training_stop": true
|
| 331 |
+
},
|
| 332 |
+
"attributes": {}
|
| 333 |
+
}
|
| 334 |
+
},
|
| 335 |
+
"total_flos": 2.8870417005060096e+16,
|
| 336 |
+
"train_batch_size": 1,
|
| 337 |
+
"trial_name": null,
|
| 338 |
+
"trial_params": null
|
| 339 |
+
}
|
checkpoint-500/adapter_config.json
ADDED
|
@@ -0,0 +1,42 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"alpha_pattern": {},
|
| 3 |
+
"auto_mapping": null,
|
| 4 |
+
"base_model_name_or_path": "/home/support/llm/Ministral-8B-Instruct-2410/",
|
| 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": 16,
|
| 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 |
+
"qalora_group_size": 16,
|
| 24 |
+
"r": 64,
|
| 25 |
+
"rank_pattern": {},
|
| 26 |
+
"revision": null,
|
| 27 |
+
"target_modules": [
|
| 28 |
+
"k_proj",
|
| 29 |
+
"o_proj",
|
| 30 |
+
"down_proj",
|
| 31 |
+
"up_proj",
|
| 32 |
+
"q_proj",
|
| 33 |
+
"v_proj",
|
| 34 |
+
"gate_proj"
|
| 35 |
+
],
|
| 36 |
+
"target_parameters": null,
|
| 37 |
+
"task_type": "CAUSAL_LM",
|
| 38 |
+
"trainable_token_indices": null,
|
| 39 |
+
"use_dora": false,
|
| 40 |
+
"use_qalora": false,
|
| 41 |
+
"use_rslora": false
|
| 42 |
+
}
|
checkpoint-500/special_tokens_map.json
ADDED
|
@@ -0,0 +1,24 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"content": "<s>",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": false,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"eos_token": {
|
| 10 |
+
"content": "</s>",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": false,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"pad_token": "</s>",
|
| 17 |
+
"unk_token": {
|
| 18 |
+
"content": "<unk>",
|
| 19 |
+
"lstrip": false,
|
| 20 |
+
"normalized": false,
|
| 21 |
+
"rstrip": false,
|
| 22 |
+
"single_word": false
|
| 23 |
+
}
|
| 24 |
+
}
|
checkpoint-500/tokenizer.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d7edbeaf20dd7f571b5dd1c54d9ace4f9b6299127cc7ba2afb14a6d51a4a79a4
|
| 3 |
+
size 17078136
|
checkpoint-500/tokenizer_config.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
checkpoint-500/trainer_state.json
ADDED
|
@@ -0,0 +1,644 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 150,
|
| 3 |
+
"best_metric": 0.5928216576576233,
|
| 4 |
+
"best_model_checkpoint": "./python-code-completion-ministral-8B/checkpoint-100",
|
| 5 |
+
"epoch": 3.9098039215686273,
|
| 6 |
+
"eval_steps": 50,
|
| 7 |
+
"global_step": 500,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"entropy": 1.5072221472859382,
|
| 14 |
+
"epoch": 0.0784313725490196,
|
| 15 |
+
"grad_norm": 0.355531245470047,
|
| 16 |
+
"learning_rate": 6e-05,
|
| 17 |
+
"loss": 1.7279,
|
| 18 |
+
"mean_token_accuracy": 0.6269892059266567,
|
| 19 |
+
"num_tokens": 22720.0,
|
| 20 |
+
"step": 10
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"entropy": 1.156855022162199,
|
| 24 |
+
"epoch": 0.1568627450980392,
|
| 25 |
+
"grad_norm": 0.4324820339679718,
|
| 26 |
+
"learning_rate": 9.998321768134102e-05,
|
| 27 |
+
"loss": 1.027,
|
| 28 |
+
"mean_token_accuracy": 0.7633639872074127,
|
| 29 |
+
"num_tokens": 46930.0,
|
| 30 |
+
"step": 20
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"entropy": 0.724648866429925,
|
| 34 |
+
"epoch": 0.23529411764705882,
|
| 35 |
+
"grad_norm": 0.28367722034454346,
|
| 36 |
+
"learning_rate": 9.979454595377594e-05,
|
| 37 |
+
"loss": 0.7082,
|
| 38 |
+
"mean_token_accuracy": 0.8542310416698455,
|
| 39 |
+
"num_tokens": 73605.0,
|
| 40 |
+
"step": 30
|
| 41 |
+
},
|
| 42 |
+
{
|
| 43 |
+
"entropy": 0.6228405788540841,
|
| 44 |
+
"epoch": 0.3137254901960784,
|
| 45 |
+
"grad_norm": 0.2075912207365036,
|
| 46 |
+
"learning_rate": 9.93970185956522e-05,
|
| 47 |
+
"loss": 0.653,
|
| 48 |
+
"mean_token_accuracy": 0.8736839070916176,
|
| 49 |
+
"num_tokens": 98799.0,
|
| 50 |
+
"step": 40
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"entropy": 0.5743573283776641,
|
| 54 |
+
"epoch": 0.39215686274509803,
|
| 55 |
+
"grad_norm": 0.21785929799079895,
|
| 56 |
+
"learning_rate": 9.879230297486034e-05,
|
| 57 |
+
"loss": 0.6108,
|
| 58 |
+
"mean_token_accuracy": 0.8821258693933487,
|
| 59 |
+
"num_tokens": 121501.0,
|
| 60 |
+
"step": 50
|
| 61 |
+
},
|
| 62 |
+
{
|
| 63 |
+
"epoch": 0.39215686274509803,
|
| 64 |
+
"eval_entropy": 0.6596304257710774,
|
| 65 |
+
"eval_loss": 0.6177071332931519,
|
| 66 |
+
"eval_mean_token_accuracy": 0.8652889298068153,
|
| 67 |
+
"eval_num_tokens": 121501.0,
|
| 68 |
+
"eval_runtime": 15.4707,
|
| 69 |
+
"eval_samples_per_second": 5.817,
|
| 70 |
+
"eval_steps_per_second": 5.817,
|
| 71 |
+
"step": 50
|
| 72 |
+
},
|
| 73 |
+
{
|
| 74 |
+
"entropy": 0.6043186411261559,
|
| 75 |
+
"epoch": 0.47058823529411764,
|
| 76 |
+
"grad_norm": 0.31095993518829346,
|
| 77 |
+
"learning_rate": 9.798293547886748e-05,
|
| 78 |
+
"loss": 0.645,
|
| 79 |
+
"mean_token_accuracy": 0.8754804000258446,
|
| 80 |
+
"num_tokens": 145419.0,
|
| 81 |
+
"step": 60
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"entropy": 0.6452899970114231,
|
| 85 |
+
"epoch": 0.5490196078431373,
|
| 86 |
+
"grad_norm": 0.3375164866447449,
|
| 87 |
+
"learning_rate": 9.697231087622691e-05,
|
| 88 |
+
"loss": 0.6634,
|
| 89 |
+
"mean_token_accuracy": 0.8640451654791832,
|
| 90 |
+
"num_tokens": 169873.0,
|
| 91 |
+
"step": 70
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"entropy": 0.6401195958256721,
|
| 95 |
+
"epoch": 0.6274509803921569,
|
| 96 |
+
"grad_norm": 0.21431401371955872,
|
| 97 |
+
"learning_rate": 9.576466807773899e-05,
|
| 98 |
+
"loss": 0.6615,
|
| 99 |
+
"mean_token_accuracy": 0.8669946551322937,
|
| 100 |
+
"num_tokens": 194305.0,
|
| 101 |
+
"step": 80
|
| 102 |
+
},
|
| 103 |
+
{
|
| 104 |
+
"entropy": 0.5998968390747905,
|
| 105 |
+
"epoch": 0.7058823529411765,
|
| 106 |
+
"grad_norm": 0.22246505320072174,
|
| 107 |
+
"learning_rate": 9.436507235698612e-05,
|
| 108 |
+
"loss": 0.6405,
|
| 109 |
+
"mean_token_accuracy": 0.877622701227665,
|
| 110 |
+
"num_tokens": 217799.0,
|
| 111 |
+
"step": 90
|
| 112 |
+
},
|
| 113 |
+
{
|
| 114 |
+
"entropy": 0.5195517055690289,
|
| 115 |
+
"epoch": 0.7843137254901961,
|
| 116 |
+
"grad_norm": 0.22644591331481934,
|
| 117 |
+
"learning_rate": 9.277939410481507e-05,
|
| 118 |
+
"loss": 0.574,
|
| 119 |
+
"mean_token_accuracy": 0.8936989307403564,
|
| 120 |
+
"num_tokens": 241897.0,
|
| 121 |
+
"step": 100
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.7843137254901961,
|
| 125 |
+
"eval_entropy": 0.607519467257791,
|
| 126 |
+
"eval_loss": 0.5936470627784729,
|
| 127 |
+
"eval_mean_token_accuracy": 0.8691287775834401,
|
| 128 |
+
"eval_num_tokens": 241897.0,
|
| 129 |
+
"eval_runtime": 15.5316,
|
| 130 |
+
"eval_samples_per_second": 5.795,
|
| 131 |
+
"eval_steps_per_second": 5.795,
|
| 132 |
+
"step": 100
|
| 133 |
+
},
|
| 134 |
+
{
|
| 135 |
+
"entropy": 0.598377226293087,
|
| 136 |
+
"epoch": 0.8627450980392157,
|
| 137 |
+
"grad_norm": 0.19417309761047363,
|
| 138 |
+
"learning_rate": 9.101428420687759e-05,
|
| 139 |
+
"loss": 0.6663,
|
| 140 |
+
"mean_token_accuracy": 0.8794096082448959,
|
| 141 |
+
"num_tokens": 266761.0,
|
| 142 |
+
"step": 110
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"entropy": 0.5745990440249443,
|
| 146 |
+
"epoch": 0.9411764705882353,
|
| 147 |
+
"grad_norm": 0.30245068669319153,
|
| 148 |
+
"learning_rate": 8.907714614750473e-05,
|
| 149 |
+
"loss": 0.605,
|
| 150 |
+
"mean_token_accuracy": 0.8777685552835465,
|
| 151 |
+
"num_tokens": 291812.0,
|
| 152 |
+
"step": 120
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"entropy": 0.563015444890449,
|
| 156 |
+
"epoch": 1.0156862745098039,
|
| 157 |
+
"grad_norm": 0.24953527748584747,
|
| 158 |
+
"learning_rate": 8.697610495692054e-05,
|
| 159 |
+
"loss": 0.5665,
|
| 160 |
+
"mean_token_accuracy": 0.8813700503424594,
|
| 161 |
+
"num_tokens": 314476.0,
|
| 162 |
+
"step": 130
|
| 163 |
+
},
|
| 164 |
+
{
|
| 165 |
+
"entropy": 0.53117360714823,
|
| 166 |
+
"epoch": 1.0941176470588236,
|
| 167 |
+
"grad_norm": 0.24786317348480225,
|
| 168 |
+
"learning_rate": 8.471997313204181e-05,
|
| 169 |
+
"loss": 0.5683,
|
| 170 |
+
"mean_token_accuracy": 0.8960023656487465,
|
| 171 |
+
"num_tokens": 335438.0,
|
| 172 |
+
"step": 140
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"entropy": 0.48921620547771455,
|
| 176 |
+
"epoch": 1.1725490196078432,
|
| 177 |
+
"grad_norm": 0.33822259306907654,
|
| 178 |
+
"learning_rate": 8.231821367380335e-05,
|
| 179 |
+
"loss": 0.5322,
|
| 180 |
+
"mean_token_accuracy": 0.8975507125258446,
|
| 181 |
+
"num_tokens": 361329.0,
|
| 182 |
+
"step": 150
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 1.1725490196078432,
|
| 186 |
+
"eval_entropy": 0.5694952074852254,
|
| 187 |
+
"eval_loss": 0.5928216576576233,
|
| 188 |
+
"eval_mean_token_accuracy": 0.8691950864262051,
|
| 189 |
+
"eval_num_tokens": 361329.0,
|
| 190 |
+
"eval_runtime": 15.4486,
|
| 191 |
+
"eval_samples_per_second": 5.826,
|
| 192 |
+
"eval_steps_per_second": 5.826,
|
| 193 |
+
"step": 150
|
| 194 |
+
},
|
| 195 |
+
{
|
| 196 |
+
"entropy": 0.5561734782531857,
|
| 197 |
+
"epoch": 1.2509803921568627,
|
| 198 |
+
"grad_norm": 0.27892613410949707,
|
| 199 |
+
"learning_rate": 7.978090039604341e-05,
|
| 200 |
+
"loss": 0.5711,
|
| 201 |
+
"mean_token_accuracy": 0.8858156308531762,
|
| 202 |
+
"num_tokens": 387834.0,
|
| 203 |
+
"step": 160
|
| 204 |
+
},
|
| 205 |
+
{
|
| 206 |
+
"entropy": 0.48290685191750526,
|
| 207 |
+
"epoch": 1.3294117647058823,
|
| 208 |
+
"grad_norm": 0.2847723662853241,
|
| 209 |
+
"learning_rate": 7.711867567242768e-05,
|
| 210 |
+
"loss": 0.5477,
|
| 211 |
+
"mean_token_accuracy": 0.8980041235685349,
|
| 212 |
+
"num_tokens": 412449.0,
|
| 213 |
+
"step": 170
|
| 214 |
+
},
|
| 215 |
+
{
|
| 216 |
+
"entropy": 0.4553432559594512,
|
| 217 |
+
"epoch": 1.4078431372549018,
|
| 218 |
+
"grad_norm": 0.35990357398986816,
|
| 219 |
+
"learning_rate": 7.434270579863548e-05,
|
| 220 |
+
"loss": 0.4953,
|
| 221 |
+
"mean_token_accuracy": 0.9054195120930671,
|
| 222 |
+
"num_tokens": 434822.0,
|
| 223 |
+
"step": 180
|
| 224 |
+
},
|
| 225 |
+
{
|
| 226 |
+
"entropy": 0.4999146230518818,
|
| 227 |
+
"epoch": 1.4862745098039216,
|
| 228 |
+
"grad_norm": 0.23805052042007446,
|
| 229 |
+
"learning_rate": 7.146463415703531e-05,
|
| 230 |
+
"loss": 0.5029,
|
| 231 |
+
"mean_token_accuracy": 0.8930570259690285,
|
| 232 |
+
"num_tokens": 461303.0,
|
| 233 |
+
"step": 190
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"entropy": 0.49880330096930264,
|
| 237 |
+
"epoch": 1.5647058823529412,
|
| 238 |
+
"grad_norm": 0.2987266480922699,
|
| 239 |
+
"learning_rate": 6.84965323802926e-05,
|
| 240 |
+
"loss": 0.5443,
|
| 241 |
+
"mean_token_accuracy": 0.8915592297911644,
|
| 242 |
+
"num_tokens": 487363.0,
|
| 243 |
+
"step": 200
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"epoch": 1.5647058823529412,
|
| 247 |
+
"eval_entropy": 0.5494673619667689,
|
| 248 |
+
"eval_loss": 0.5963289737701416,
|
| 249 |
+
"eval_mean_token_accuracy": 0.8694895923137664,
|
| 250 |
+
"eval_num_tokens": 487363.0,
|
| 251 |
+
"eval_runtime": 15.5114,
|
| 252 |
+
"eval_samples_per_second": 5.802,
|
| 253 |
+
"eval_steps_per_second": 5.802,
|
| 254 |
+
"step": 200
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"entropy": 0.49026294946670534,
|
| 258 |
+
"epoch": 1.6431372549019607,
|
| 259 |
+
"grad_norm": 0.2591567039489746,
|
| 260 |
+
"learning_rate": 6.545084971874738e-05,
|
| 261 |
+
"loss": 0.5028,
|
| 262 |
+
"mean_token_accuracy": 0.8963919445872307,
|
| 263 |
+
"num_tokens": 512437.0,
|
| 264 |
+
"step": 210
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"entropy": 0.5070990648120641,
|
| 268 |
+
"epoch": 1.7215686274509805,
|
| 269 |
+
"grad_norm": 0.3534194231033325,
|
| 270 |
+
"learning_rate": 6.23403608239317e-05,
|
| 271 |
+
"loss": 0.5434,
|
| 272 |
+
"mean_token_accuracy": 0.8927687123417855,
|
| 273 |
+
"num_tokens": 534856.0,
|
| 274 |
+
"step": 220
|
| 275 |
+
},
|
| 276 |
+
{
|
| 277 |
+
"entropy": 0.5167016156017781,
|
| 278 |
+
"epoch": 1.8,
|
| 279 |
+
"grad_norm": 0.26480528712272644,
|
| 280 |
+
"learning_rate": 5.9178112167241805e-05,
|
| 281 |
+
"loss": 0.5265,
|
| 282 |
+
"mean_token_accuracy": 0.8922785848379136,
|
| 283 |
+
"num_tokens": 560263.0,
|
| 284 |
+
"step": 230
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"entropy": 0.4376767996698618,
|
| 288 |
+
"epoch": 1.8784313725490196,
|
| 289 |
+
"grad_norm": 0.2970808148384094,
|
| 290 |
+
"learning_rate": 5.597736731850295e-05,
|
| 291 |
+
"loss": 0.4587,
|
| 292 |
+
"mean_token_accuracy": 0.9108576759696007,
|
| 293 |
+
"num_tokens": 582717.0,
|
| 294 |
+
"step": 240
|
| 295 |
+
},
|
| 296 |
+
{
|
| 297 |
+
"entropy": 0.5208950355648995,
|
| 298 |
+
"epoch": 1.9568627450980394,
|
| 299 |
+
"grad_norm": 0.27791768312454224,
|
| 300 |
+
"learning_rate": 5.2751551313948246e-05,
|
| 301 |
+
"loss": 0.5336,
|
| 302 |
+
"mean_token_accuracy": 0.8880411878228187,
|
| 303 |
+
"num_tokens": 608651.0,
|
| 304 |
+
"step": 250
|
| 305 |
+
},
|
| 306 |
+
{
|
| 307 |
+
"epoch": 1.9568627450980394,
|
| 308 |
+
"eval_entropy": 0.5718106091022491,
|
| 309 |
+
"eval_loss": 0.5953241586685181,
|
| 310 |
+
"eval_mean_token_accuracy": 0.8688194268279605,
|
| 311 |
+
"eval_num_tokens": 608651.0,
|
| 312 |
+
"eval_runtime": 15.4979,
|
| 313 |
+
"eval_samples_per_second": 5.807,
|
| 314 |
+
"eval_steps_per_second": 5.807,
|
| 315 |
+
"step": 250
|
| 316 |
+
},
|
| 317 |
+
{
|
| 318 |
+
"entropy": 0.45974215472999375,
|
| 319 |
+
"epoch": 2.0313725490196077,
|
| 320 |
+
"grad_norm": 0.32541951537132263,
|
| 321 |
+
"learning_rate": 4.9514194346951144e-05,
|
| 322 |
+
"loss": 0.4736,
|
| 323 |
+
"mean_token_accuracy": 0.9053617276643452,
|
| 324 |
+
"num_tokens": 629040.0,
|
| 325 |
+
"step": 260
|
| 326 |
+
},
|
| 327 |
+
{
|
| 328 |
+
"entropy": 0.4414994928985834,
|
| 329 |
+
"epoch": 2.1098039215686275,
|
| 330 |
+
"grad_norm": 0.372414231300354,
|
| 331 |
+
"learning_rate": 4.627887501769231e-05,
|
| 332 |
+
"loss": 0.4428,
|
| 333 |
+
"mean_token_accuracy": 0.9085879683494568,
|
| 334 |
+
"num_tokens": 654385.0,
|
| 335 |
+
"step": 270
|
| 336 |
+
},
|
| 337 |
+
{
|
| 338 |
+
"entropy": 0.3868284210562706,
|
| 339 |
+
"epoch": 2.1882352941176473,
|
| 340 |
+
"grad_norm": 0.27554434537887573,
|
| 341 |
+
"learning_rate": 4.305916337979168e-05,
|
| 342 |
+
"loss": 0.4041,
|
| 343 |
+
"mean_token_accuracy": 0.9162766322493553,
|
| 344 |
+
"num_tokens": 677869.0,
|
| 345 |
+
"step": 280
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"entropy": 0.48472463674843314,
|
| 349 |
+
"epoch": 2.2666666666666666,
|
| 350 |
+
"grad_norm": 0.3065558969974518,
|
| 351 |
+
"learning_rate": 3.9868564022787505e-05,
|
| 352 |
+
"loss": 0.4702,
|
| 353 |
+
"mean_token_accuracy": 0.9028230115771294,
|
| 354 |
+
"num_tokens": 705954.0,
|
| 355 |
+
"step": 290
|
| 356 |
+
},
|
| 357 |
+
{
|
| 358 |
+
"entropy": 0.35362908374518154,
|
| 359 |
+
"epoch": 2.3450980392156864,
|
| 360 |
+
"grad_norm": 0.42765629291534424,
|
| 361 |
+
"learning_rate": 3.672045942919474e-05,
|
| 362 |
+
"loss": 0.3858,
|
| 363 |
+
"mean_token_accuracy": 0.9247043713927269,
|
| 364 |
+
"num_tokens": 729855.0,
|
| 365 |
+
"step": 300
|
| 366 |
+
},
|
| 367 |
+
{
|
| 368 |
+
"epoch": 2.3450980392156864,
|
| 369 |
+
"eval_entropy": 0.4895121798747116,
|
| 370 |
+
"eval_loss": 0.633377194404602,
|
| 371 |
+
"eval_mean_token_accuracy": 0.8656294153796302,
|
| 372 |
+
"eval_num_tokens": 729855.0,
|
| 373 |
+
"eval_runtime": 15.4754,
|
| 374 |
+
"eval_samples_per_second": 5.816,
|
| 375 |
+
"eval_steps_per_second": 5.816,
|
| 376 |
+
"step": 300
|
| 377 |
+
},
|
| 378 |
+
{
|
| 379 |
+
"entropy": 0.3393667917698622,
|
| 380 |
+
"epoch": 2.4235294117647057,
|
| 381 |
+
"grad_norm": 0.3438664972782135,
|
| 382 |
+
"learning_rate": 3.362805384372267e-05,
|
| 383 |
+
"loss": 0.3374,
|
| 384 |
+
"mean_token_accuracy": 0.9306267336010933,
|
| 385 |
+
"num_tokens": 750942.0,
|
| 386 |
+
"step": 310
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"entropy": 0.3776489123702049,
|
| 390 |
+
"epoch": 2.5019607843137255,
|
| 391 |
+
"grad_norm": 0.3191138207912445,
|
| 392 |
+
"learning_rate": 3.0604317890083676e-05,
|
| 393 |
+
"loss": 0.3675,
|
| 394 |
+
"mean_token_accuracy": 0.9214938268065452,
|
| 395 |
+
"num_tokens": 774956.0,
|
| 396 |
+
"step": 320
|
| 397 |
+
},
|
| 398 |
+
{
|
| 399 |
+
"entropy": 0.39363621212542055,
|
| 400 |
+
"epoch": 2.5803921568627453,
|
| 401 |
+
"grad_norm": 0.4947509169578552,
|
| 402 |
+
"learning_rate": 2.7661934167689884e-05,
|
| 403 |
+
"loss": 0.4188,
|
| 404 |
+
"mean_token_accuracy": 0.916774469614029,
|
| 405 |
+
"num_tokens": 800996.0,
|
| 406 |
+
"step": 330
|
| 407 |
+
},
|
| 408 |
+
{
|
| 409 |
+
"entropy": 0.3987773071974516,
|
| 410 |
+
"epoch": 2.6588235294117646,
|
| 411 |
+
"grad_norm": 0.4273108243942261,
|
| 412 |
+
"learning_rate": 2.481324405642369e-05,
|
| 413 |
+
"loss": 0.4063,
|
| 414 |
+
"mean_token_accuracy": 0.9157314419746398,
|
| 415 |
+
"num_tokens": 826362.0,
|
| 416 |
+
"step": 340
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"entropy": 0.4490684296935797,
|
| 420 |
+
"epoch": 2.7372549019607844,
|
| 421 |
+
"grad_norm": 0.47922441363334656,
|
| 422 |
+
"learning_rate": 2.207019595260154e-05,
|
| 423 |
+
"loss": 0.4576,
|
| 424 |
+
"mean_token_accuracy": 0.9035186737775802,
|
| 425 |
+
"num_tokens": 853427.0,
|
| 426 |
+
"step": 350
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 2.7372549019607844,
|
| 430 |
+
"eval_entropy": 0.5091634632812606,
|
| 431 |
+
"eval_loss": 0.6345335245132446,
|
| 432 |
+
"eval_mean_token_accuracy": 0.8654518723487854,
|
| 433 |
+
"eval_num_tokens": 853427.0,
|
| 434 |
+
"eval_runtime": 15.4411,
|
| 435 |
+
"eval_samples_per_second": 5.829,
|
| 436 |
+
"eval_steps_per_second": 5.829,
|
| 437 |
+
"step": 350
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"entropy": 0.3651732888072729,
|
| 441 |
+
"epoch": 2.8156862745098037,
|
| 442 |
+
"grad_norm": 0.4229300618171692,
|
| 443 |
+
"learning_rate": 1.9444295153247487e-05,
|
| 444 |
+
"loss": 0.3517,
|
| 445 |
+
"mean_token_accuracy": 0.9295069724321365,
|
| 446 |
+
"num_tokens": 876031.0,
|
| 447 |
+
"step": 360
|
| 448 |
+
},
|
| 449 |
+
{
|
| 450 |
+
"entropy": 0.37069291807711124,
|
| 451 |
+
"epoch": 2.8941176470588235,
|
| 452 |
+
"grad_norm": 0.383728951215744,
|
| 453 |
+
"learning_rate": 1.694655559887914e-05,
|
| 454 |
+
"loss": 0.3875,
|
| 455 |
+
"mean_token_accuracy": 0.9240497961640358,
|
| 456 |
+
"num_tokens": 898756.0,
|
| 457 |
+
"step": 370
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"entropy": 0.3721807461231947,
|
| 461 |
+
"epoch": 2.9725490196078432,
|
| 462 |
+
"grad_norm": 0.40181341767311096,
|
| 463 |
+
"learning_rate": 1.4587453677213348e-05,
|
| 464 |
+
"loss": 0.3983,
|
| 465 |
+
"mean_token_accuracy": 0.9216682553291321,
|
| 466 |
+
"num_tokens": 921618.0,
|
| 467 |
+
"step": 380
|
| 468 |
+
},
|
| 469 |
+
{
|
| 470 |
+
"entropy": 0.4022895936903201,
|
| 471 |
+
"epoch": 3.0470588235294116,
|
| 472 |
+
"grad_norm": 0.32404816150665283,
|
| 473 |
+
"learning_rate": 1.2376884281555484e-05,
|
| 474 |
+
"loss": 0.3566,
|
| 475 |
+
"mean_token_accuracy": 0.9227047578284615,
|
| 476 |
+
"num_tokens": 948594.0,
|
| 477 |
+
"step": 390
|
| 478 |
+
},
|
| 479 |
+
{
|
| 480 |
+
"entropy": 0.3646719891577959,
|
| 481 |
+
"epoch": 3.1254901960784314,
|
| 482 |
+
"grad_norm": 0.3212927281856537,
|
| 483 |
+
"learning_rate": 1.0324119308178166e-05,
|
| 484 |
+
"loss": 0.3489,
|
| 485 |
+
"mean_token_accuracy": 0.9308759719133377,
|
| 486 |
+
"num_tokens": 972074.0,
|
| 487 |
+
"step": 400
|
| 488 |
+
},
|
| 489 |
+
{
|
| 490 |
+
"epoch": 3.1254901960784314,
|
| 491 |
+
"eval_entropy": 0.47637585144903927,
|
| 492 |
+
"eval_loss": 0.661882758140564,
|
| 493 |
+
"eval_mean_token_accuracy": 0.8632458435164557,
|
| 494 |
+
"eval_num_tokens": 972074.0,
|
| 495 |
+
"eval_runtime": 15.4948,
|
| 496 |
+
"eval_samples_per_second": 5.808,
|
| 497 |
+
"eval_steps_per_second": 5.808,
|
| 498 |
+
"step": 400
|
| 499 |
+
},
|
| 500 |
+
{
|
| 501 |
+
"entropy": 0.30484746433794496,
|
| 502 |
+
"epoch": 3.203921568627451,
|
| 503 |
+
"grad_norm": 0.3374861776828766,
|
| 504 |
+
"learning_rate": 8.437768766765974e-06,
|
| 505 |
+
"loss": 0.3161,
|
| 506 |
+
"mean_token_accuracy": 0.9372083112597466,
|
| 507 |
+
"num_tokens": 994846.0,
|
| 508 |
+
"step": 410
|
| 509 |
+
},
|
| 510 |
+
{
|
| 511 |
+
"entropy": 0.35498267617076634,
|
| 512 |
+
"epoch": 3.2823529411764705,
|
| 513 |
+
"grad_norm": 0.36716318130493164,
|
| 514 |
+
"learning_rate": 6.725744667042777e-06,
|
| 515 |
+
"loss": 0.3387,
|
| 516 |
+
"mean_token_accuracy": 0.9290374144911766,
|
| 517 |
+
"num_tokens": 1019726.0,
|
| 518 |
+
"step": 420
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"entropy": 0.3321664972230792,
|
| 522 |
+
"epoch": 3.3607843137254902,
|
| 523 |
+
"grad_norm": 0.4375359117984772,
|
| 524 |
+
"learning_rate": 5.195227833053274e-06,
|
| 525 |
+
"loss": 0.323,
|
| 526 |
+
"mean_token_accuracy": 0.9353578373789787,
|
| 527 |
+
"num_tokens": 1044999.0,
|
| 528 |
+
"step": 430
|
| 529 |
+
},
|
| 530 |
+
{
|
| 531 |
+
"entropy": 0.33164924252778294,
|
| 532 |
+
"epoch": 3.43921568627451,
|
| 533 |
+
"grad_norm": 0.48313814401626587,
|
| 534 |
+
"learning_rate": 3.852637784291424e-06,
|
| 535 |
+
"loss": 0.346,
|
| 536 |
+
"mean_token_accuracy": 0.9326006546616554,
|
| 537 |
+
"num_tokens": 1068850.0,
|
| 538 |
+
"step": 440
|
| 539 |
+
},
|
| 540 |
+
{
|
| 541 |
+
"entropy": 0.29590620547533036,
|
| 542 |
+
"epoch": 3.5176470588235293,
|
| 543 |
+
"grad_norm": 0.5358214378356934,
|
| 544 |
+
"learning_rate": 2.703605810004772e-06,
|
| 545 |
+
"loss": 0.2704,
|
| 546 |
+
"mean_token_accuracy": 0.9426496475934982,
|
| 547 |
+
"num_tokens": 1092853.0,
|
| 548 |
+
"step": 450
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 3.5176470588235293,
|
| 552 |
+
"eval_entropy": 0.46879171944326825,
|
| 553 |
+
"eval_loss": 0.6711850762367249,
|
| 554 |
+
"eval_mean_token_accuracy": 0.8622583528359731,
|
| 555 |
+
"eval_num_tokens": 1092853.0,
|
| 556 |
+
"eval_runtime": 15.462,
|
| 557 |
+
"eval_samples_per_second": 5.821,
|
| 558 |
+
"eval_steps_per_second": 5.821,
|
| 559 |
+
"step": 450
|
| 560 |
+
},
|
| 561 |
+
{
|
| 562 |
+
"entropy": 0.3179448990151286,
|
| 563 |
+
"epoch": 3.596078431372549,
|
| 564 |
+
"grad_norm": 0.4179091155529022,
|
| 565 |
+
"learning_rate": 1.7529513496103323e-06,
|
| 566 |
+
"loss": 0.3118,
|
| 567 |
+
"mean_token_accuracy": 0.9393924847245216,
|
| 568 |
+
"num_tokens": 1115370.0,
|
| 569 |
+
"step": 460
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"entropy": 0.31651122458279135,
|
| 573 |
+
"epoch": 3.674509803921569,
|
| 574 |
+
"grad_norm": 0.4960501492023468,
|
| 575 |
+
"learning_rate": 1.004661778290783e-06,
|
| 576 |
+
"loss": 0.3162,
|
| 577 |
+
"mean_token_accuracy": 0.9382667735219001,
|
| 578 |
+
"num_tokens": 1137241.0,
|
| 579 |
+
"step": 470
|
| 580 |
+
},
|
| 581 |
+
{
|
| 582 |
+
"entropy": 0.3232851158827543,
|
| 583 |
+
"epoch": 3.7529411764705882,
|
| 584 |
+
"grad_norm": 0.39605653285980225,
|
| 585 |
+
"learning_rate": 4.6187568255726124e-07,
|
| 586 |
+
"loss": 0.3284,
|
| 587 |
+
"mean_token_accuracy": 0.9355289787054062,
|
| 588 |
+
"num_tokens": 1162001.0,
|
| 589 |
+
"step": 480
|
| 590 |
+
},
|
| 591 |
+
{
|
| 592 |
+
"entropy": 0.3265875181183219,
|
| 593 |
+
"epoch": 3.831372549019608,
|
| 594 |
+
"grad_norm": 0.6258050799369812,
|
| 595 |
+
"learning_rate": 1.268696959267679e-07,
|
| 596 |
+
"loss": 0.3199,
|
| 597 |
+
"mean_token_accuracy": 0.9376031413674355,
|
| 598 |
+
"num_tokens": 1186075.0,
|
| 599 |
+
"step": 490
|
| 600 |
+
},
|
| 601 |
+
{
|
| 602 |
+
"entropy": 0.36884093452245,
|
| 603 |
+
"epoch": 3.9098039215686273,
|
| 604 |
+
"grad_norm": 0.44404006004333496,
|
| 605 |
+
"learning_rate": 1.0489499300603278e-09,
|
| 606 |
+
"loss": 0.3421,
|
| 607 |
+
"mean_token_accuracy": 0.9266003414988517,
|
| 608 |
+
"num_tokens": 1213302.0,
|
| 609 |
+
"step": 500
|
| 610 |
+
},
|
| 611 |
+
{
|
| 612 |
+
"epoch": 3.9098039215686273,
|
| 613 |
+
"eval_entropy": 0.47118989841805564,
|
| 614 |
+
"eval_loss": 0.6726338863372803,
|
| 615 |
+
"eval_mean_token_accuracy": 0.8621971819135877,
|
| 616 |
+
"eval_num_tokens": 1213302.0,
|
| 617 |
+
"eval_runtime": 15.5109,
|
| 618 |
+
"eval_samples_per_second": 5.802,
|
| 619 |
+
"eval_steps_per_second": 5.802,
|
| 620 |
+
"step": 500
|
| 621 |
+
}
|
| 622 |
+
],
|
| 623 |
+
"logging_steps": 10,
|
| 624 |
+
"max_steps": 500,
|
| 625 |
+
"num_input_tokens_seen": 0,
|
| 626 |
+
"num_train_epochs": 4,
|
| 627 |
+
"save_steps": 100,
|
| 628 |
+
"stateful_callbacks": {
|
| 629 |
+
"TrainerControl": {
|
| 630 |
+
"args": {
|
| 631 |
+
"should_epoch_stop": false,
|
| 632 |
+
"should_evaluate": false,
|
| 633 |
+
"should_log": false,
|
| 634 |
+
"should_save": true,
|
| 635 |
+
"should_training_stop": true
|
| 636 |
+
},
|
| 637 |
+
"attributes": {}
|
| 638 |
+
}
|
| 639 |
+
},
|
| 640 |
+
"total_flos": 5.574534419069338e+16,
|
| 641 |
+
"train_batch_size": 1,
|
| 642 |
+
"trial_name": null,
|
| 643 |
+
"trial_params": null
|
| 644 |
+
}
|
special_tokens_map.json
ADDED
|
@@ -0,0 +1,24 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"content": "<s>",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": false,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"eos_token": {
|
| 10 |
+
"content": "</s>",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": false,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"pad_token": "</s>",
|
| 17 |
+
"unk_token": {
|
| 18 |
+
"content": "<unk>",
|
| 19 |
+
"lstrip": false,
|
| 20 |
+
"normalized": false,
|
| 21 |
+
"rstrip": false,
|
| 22 |
+
"single_word": false
|
| 23 |
+
}
|
| 24 |
+
}
|
tokenizer.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d7edbeaf20dd7f571b5dd1c54d9ace4f9b6299127cc7ba2afb14a6d51a4a79a4
|
| 3 |
+
size 17078136
|
tokenizer_config.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|