Upload 3 files
Browse files
dclm-1.0-baseline/config.json
ADDED
|
@@ -0,0 +1,32 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"LlamaForCausalLM"
|
| 4 |
+
],
|
| 5 |
+
"attention_bias": false,
|
| 6 |
+
"attention_dropout": 0.0,
|
| 7 |
+
"bos_token_id": 1,
|
| 8 |
+
"dtype": "float32",
|
| 9 |
+
"eos_token_id": 2,
|
| 10 |
+
"head_dim": 32,
|
| 11 |
+
"hidden_act": "silu",
|
| 12 |
+
"hidden_size": 128,
|
| 13 |
+
"initializer_range": 0.02,
|
| 14 |
+
"intermediate_size": 344,
|
| 15 |
+
"max_position_embeddings": 256,
|
| 16 |
+
"mlp_bias": false,
|
| 17 |
+
"model_type": "llama",
|
| 18 |
+
"num_attention_heads": 4,
|
| 19 |
+
"num_hidden_layers": 8,
|
| 20 |
+
"num_key_value_heads": 2,
|
| 21 |
+
"pad_token_id": null,
|
| 22 |
+
"pretraining_tp": 1,
|
| 23 |
+
"rms_norm_eps": 1e-06,
|
| 24 |
+
"rope_parameters": {
|
| 25 |
+
"rope_theta": 2500.0,
|
| 26 |
+
"rope_type": "default"
|
| 27 |
+
},
|
| 28 |
+
"tie_word_embeddings": true,
|
| 29 |
+
"transformers_version": "5.13.1",
|
| 30 |
+
"use_cache": false,
|
| 31 |
+
"vocab_size": 2564
|
| 32 |
+
}
|
dclm-1.0-baseline/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:844838fb74d820e76b633a69f1dc3efb40163a448aa473556d8c6d4b1d21c0ae
|
| 3 |
+
size 7129288
|
dclm-1.0-baseline/trainer_state.json
ADDED
|
@@ -0,0 +1,242 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": null,
|
| 3 |
+
"best_metric": null,
|
| 4 |
+
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 1.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 2428,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.0411946446961895,
|
| 14 |
+
"grad_norm": 0.467864453792572,
|
| 15 |
+
"learning_rate": 0.003,
|
| 16 |
+
"loss": 6.407201538085937,
|
| 17 |
+
"step": 100
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.082389289392379,
|
| 21 |
+
"grad_norm": 0.4472709596157074,
|
| 22 |
+
"learning_rate": 0.003,
|
| 23 |
+
"loss": 5.041720581054688,
|
| 24 |
+
"step": 200
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.12358393408856849,
|
| 28 |
+
"grad_norm": 0.48611021041870117,
|
| 29 |
+
"learning_rate": 0.003,
|
| 30 |
+
"loss": 4.333825988769531,
|
| 31 |
+
"step": 300
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.164778578784758,
|
| 35 |
+
"grad_norm": 0.406901091337204,
|
| 36 |
+
"learning_rate": 0.003,
|
| 37 |
+
"loss": 4.042262573242187,
|
| 38 |
+
"step": 400
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.2059732234809475,
|
| 42 |
+
"grad_norm": 0.33796098828315735,
|
| 43 |
+
"learning_rate": 0.003,
|
| 44 |
+
"loss": 3.9043743896484373,
|
| 45 |
+
"step": 500
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.2059732234809475,
|
| 49 |
+
"eval_loss": 3.865790843963623,
|
| 50 |
+
"eval_runtime": 8.0265,
|
| 51 |
+
"eval_samples_per_second": 399.925,
|
| 52 |
+
"eval_steps_per_second": 50.084,
|
| 53 |
+
"step": 500
|
| 54 |
+
},
|
| 55 |
+
{
|
| 56 |
+
"epoch": 0.24716786817713698,
|
| 57 |
+
"grad_norm": 0.4365800619125366,
|
| 58 |
+
"learning_rate": 0.003,
|
| 59 |
+
"loss": 3.8237649536132814,
|
| 60 |
+
"step": 600
|
| 61 |
+
},
|
| 62 |
+
{
|
| 63 |
+
"epoch": 0.2883625128733265,
|
| 64 |
+
"grad_norm": 0.28054454922676086,
|
| 65 |
+
"learning_rate": 0.003,
|
| 66 |
+
"loss": 3.764461669921875,
|
| 67 |
+
"step": 700
|
| 68 |
+
},
|
| 69 |
+
{
|
| 70 |
+
"epoch": 0.329557157569516,
|
| 71 |
+
"grad_norm": 0.29116520285606384,
|
| 72 |
+
"learning_rate": 0.003,
|
| 73 |
+
"loss": 3.7118209838867187,
|
| 74 |
+
"step": 800
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.3707518022657055,
|
| 78 |
+
"grad_norm": 0.3120614290237427,
|
| 79 |
+
"learning_rate": 0.003,
|
| 80 |
+
"loss": 3.6665850830078126,
|
| 81 |
+
"step": 900
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"epoch": 0.411946446961895,
|
| 85 |
+
"grad_norm": 0.2823605239391327,
|
| 86 |
+
"learning_rate": 0.003,
|
| 87 |
+
"loss": 3.6290557861328123,
|
| 88 |
+
"step": 1000
|
| 89 |
+
},
|
| 90 |
+
{
|
| 91 |
+
"epoch": 0.411946446961895,
|
| 92 |
+
"eval_loss": 3.618884801864624,
|
| 93 |
+
"eval_runtime": 7.522,
|
| 94 |
+
"eval_samples_per_second": 426.746,
|
| 95 |
+
"eval_steps_per_second": 53.443,
|
| 96 |
+
"step": 1000
|
| 97 |
+
},
|
| 98 |
+
{
|
| 99 |
+
"epoch": 0.45314109165808447,
|
| 100 |
+
"grad_norm": 0.32250627875328064,
|
| 101 |
+
"learning_rate": 0.003,
|
| 102 |
+
"loss": 3.5978604125976563,
|
| 103 |
+
"step": 1100
|
| 104 |
+
},
|
| 105 |
+
{
|
| 106 |
+
"epoch": 0.49433573635427397,
|
| 107 |
+
"grad_norm": 0.25370901823043823,
|
| 108 |
+
"learning_rate": 0.003,
|
| 109 |
+
"loss": 3.5753305053710935,
|
| 110 |
+
"step": 1200
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.5355303810504635,
|
| 114 |
+
"grad_norm": 0.27662038803100586,
|
| 115 |
+
"learning_rate": 0.003,
|
| 116 |
+
"loss": 3.553804626464844,
|
| 117 |
+
"step": 1300
|
| 118 |
+
},
|
| 119 |
+
{
|
| 120 |
+
"epoch": 0.576725025746653,
|
| 121 |
+
"grad_norm": 0.2849913537502289,
|
| 122 |
+
"learning_rate": 0.003,
|
| 123 |
+
"loss": 3.5376666259765623,
|
| 124 |
+
"step": 1400
|
| 125 |
+
},
|
| 126 |
+
{
|
| 127 |
+
"epoch": 0.6179196704428425,
|
| 128 |
+
"grad_norm": 0.2572985291481018,
|
| 129 |
+
"learning_rate": 0.003,
|
| 130 |
+
"loss": 3.5211285400390624,
|
| 131 |
+
"step": 1500
|
| 132 |
+
},
|
| 133 |
+
{
|
| 134 |
+
"epoch": 0.6179196704428425,
|
| 135 |
+
"eval_loss": 3.521266460418701,
|
| 136 |
+
"eval_runtime": 8.3868,
|
| 137 |
+
"eval_samples_per_second": 382.744,
|
| 138 |
+
"eval_steps_per_second": 47.932,
|
| 139 |
+
"step": 1500
|
| 140 |
+
},
|
| 141 |
+
{
|
| 142 |
+
"epoch": 0.659114315139032,
|
| 143 |
+
"grad_norm": 0.21017690002918243,
|
| 144 |
+
"learning_rate": 0.003,
|
| 145 |
+
"loss": 3.5094705200195313,
|
| 146 |
+
"step": 1600
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 0.7003089598352215,
|
| 150 |
+
"grad_norm": 0.2542063891887665,
|
| 151 |
+
"learning_rate": 0.003,
|
| 152 |
+
"loss": 3.4962704467773436,
|
| 153 |
+
"step": 1700
|
| 154 |
+
},
|
| 155 |
+
{
|
| 156 |
+
"epoch": 0.741503604531411,
|
| 157 |
+
"grad_norm": 0.22395335137844086,
|
| 158 |
+
"learning_rate": 0.003,
|
| 159 |
+
"loss": 3.485750427246094,
|
| 160 |
+
"step": 1800
|
| 161 |
+
},
|
| 162 |
+
{
|
| 163 |
+
"epoch": 0.7826982492276005,
|
| 164 |
+
"grad_norm": 0.27532121539115906,
|
| 165 |
+
"learning_rate": 0.003,
|
| 166 |
+
"loss": 3.4784066772460935,
|
| 167 |
+
"step": 1900
|
| 168 |
+
},
|
| 169 |
+
{
|
| 170 |
+
"epoch": 0.82389289392379,
|
| 171 |
+
"grad_norm": 0.22537492215633392,
|
| 172 |
+
"learning_rate": 0.002899325557098001,
|
| 173 |
+
"loss": 3.4674105834960938,
|
| 174 |
+
"step": 2000
|
| 175 |
+
},
|
| 176 |
+
{
|
| 177 |
+
"epoch": 0.82389289392379,
|
| 178 |
+
"eval_loss": 3.4714837074279785,
|
| 179 |
+
"eval_runtime": 7.2122,
|
| 180 |
+
"eval_samples_per_second": 445.079,
|
| 181 |
+
"eval_steps_per_second": 55.739,
|
| 182 |
+
"step": 2000
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 0.8650875386199794,
|
| 186 |
+
"grad_norm": 0.18755502998828888,
|
| 187 |
+
"learning_rate": 0.0022915870802995226,
|
| 188 |
+
"loss": 3.4546932983398437,
|
| 189 |
+
"step": 2100
|
| 190 |
+
},
|
| 191 |
+
{
|
| 192 |
+
"epoch": 0.9062821833161689,
|
| 193 |
+
"grad_norm": 0.17484472692012787,
|
| 194 |
+
"learning_rate": 0.0013644373889485982,
|
| 195 |
+
"loss": 3.42761962890625,
|
| 196 |
+
"step": 2200
|
| 197 |
+
},
|
| 198 |
+
{
|
| 199 |
+
"epoch": 0.9474768280123584,
|
| 200 |
+
"grad_norm": 0.1623256355524063,
|
| 201 |
+
"learning_rate": 0.0004919882091667163,
|
| 202 |
+
"loss": 3.4039578247070312,
|
| 203 |
+
"step": 2300
|
| 204 |
+
},
|
| 205 |
+
{
|
| 206 |
+
"epoch": 0.9886714727085479,
|
| 207 |
+
"grad_norm": 0.1342345029115677,
|
| 208 |
+
"learning_rate": 2.6279207952956684e-05,
|
| 209 |
+
"loss": 3.3823727416992186,
|
| 210 |
+
"step": 2400
|
| 211 |
+
},
|
| 212 |
+
{
|
| 213 |
+
"epoch": 1.0,
|
| 214 |
+
"eval_loss": 3.383687973022461,
|
| 215 |
+
"eval_runtime": 11.5434,
|
| 216 |
+
"eval_samples_per_second": 278.081,
|
| 217 |
+
"eval_steps_per_second": 34.825,
|
| 218 |
+
"step": 2428
|
| 219 |
+
}
|
| 220 |
+
],
|
| 221 |
+
"logging_steps": 100,
|
| 222 |
+
"max_steps": 2428,
|
| 223 |
+
"num_input_tokens_seen": 0,
|
| 224 |
+
"num_train_epochs": 1,
|
| 225 |
+
"save_steps": 500,
|
| 226 |
+
"stateful_callbacks": {
|
| 227 |
+
"TrainerControl": {
|
| 228 |
+
"args": {
|
| 229 |
+
"should_epoch_stop": false,
|
| 230 |
+
"should_evaluate": false,
|
| 231 |
+
"should_log": false,
|
| 232 |
+
"should_save": true,
|
| 233 |
+
"should_training_stop": true
|
| 234 |
+
},
|
| 235 |
+
"attributes": {}
|
| 236 |
+
}
|
| 237 |
+
},
|
| 238 |
+
"total_flos": 4331665489920000.0,
|
| 239 |
+
"train_batch_size": 400,
|
| 240 |
+
"trial_name": null,
|
| 241 |
+
"trial_params": null
|
| 242 |
+
}
|