Upload folder using huggingface_hub
Browse files- config.json +64 -0
- experiment_cfg/metadata.json +363 -0
- model-00001-of-00002.safetensors +3 -0
- model-00002-of-00002.safetensors +3 -0
- model.safetensors.index.json +0 -0
- trainer_state.json +393 -0
- training_args.bin +3 -0
config.json
ADDED
|
@@ -0,0 +1,64 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"action_dim": 32,
|
| 3 |
+
"action_head_cfg": {
|
| 4 |
+
"action_dim": 32,
|
| 5 |
+
"action_horizon": 16,
|
| 6 |
+
"add_pos_embed": true,
|
| 7 |
+
"backbone_embedding_dim": 2048,
|
| 8 |
+
"diffusion_model_cfg": {
|
| 9 |
+
"attention_head_dim": 48,
|
| 10 |
+
"cross_attention_dim": 2048,
|
| 11 |
+
"dropout": 0.2,
|
| 12 |
+
"final_dropout": true,
|
| 13 |
+
"interleave_self_attention": true,
|
| 14 |
+
"norm_type": "ada_norm",
|
| 15 |
+
"num_attention_heads": 32,
|
| 16 |
+
"num_layers": 16,
|
| 17 |
+
"output_dim": 1024,
|
| 18 |
+
"positional_embeddings": null
|
| 19 |
+
},
|
| 20 |
+
"hidden_size": 1024,
|
| 21 |
+
"input_embedding_dim": 1536,
|
| 22 |
+
"max_action_dim": 32,
|
| 23 |
+
"max_state_dim": 64,
|
| 24 |
+
"model_dtype": "float32",
|
| 25 |
+
"noise_beta_alpha": 1.5,
|
| 26 |
+
"noise_beta_beta": 1.0,
|
| 27 |
+
"noise_s": 0.999,
|
| 28 |
+
"num_inference_timesteps": 4,
|
| 29 |
+
"num_target_vision_tokens": 32,
|
| 30 |
+
"num_timestep_buckets": 1000,
|
| 31 |
+
"tune_diffusion_model": true,
|
| 32 |
+
"tune_projector": true,
|
| 33 |
+
"use_vlln": true,
|
| 34 |
+
"vl_self_attention_cfg": {
|
| 35 |
+
"attention_head_dim": 64,
|
| 36 |
+
"dropout": 0.2,
|
| 37 |
+
"final_dropout": true,
|
| 38 |
+
"num_attention_heads": 32,
|
| 39 |
+
"num_layers": 4,
|
| 40 |
+
"positional_embeddings": null
|
| 41 |
+
}
|
| 42 |
+
},
|
| 43 |
+
"action_horizon": 16,
|
| 44 |
+
"architectures": [
|
| 45 |
+
"GR00T_N1_5"
|
| 46 |
+
],
|
| 47 |
+
"attn_implementation": null,
|
| 48 |
+
"backbone_cfg": {
|
| 49 |
+
"eagle_path": "NVEagle/eagle_er-qwen3_1_7B-Siglip2_400M_stage1_5_128gpu_er_v7_1mlp_nops",
|
| 50 |
+
"load_bf16": false,
|
| 51 |
+
"project_to_dim": null,
|
| 52 |
+
"reproject_vision": false,
|
| 53 |
+
"select_layer": 12,
|
| 54 |
+
"tune_llm": false,
|
| 55 |
+
"tune_visual": true,
|
| 56 |
+
"use_flash_attention": true
|
| 57 |
+
},
|
| 58 |
+
"compute_dtype": "bfloat16",
|
| 59 |
+
"hidden_size": 2048,
|
| 60 |
+
"model_dtype": "float32",
|
| 61 |
+
"model_type": "gr00t_n1_5",
|
| 62 |
+
"torch_dtype": "bfloat16",
|
| 63 |
+
"transformers_version": "4.51.3"
|
| 64 |
+
}
|
experiment_cfg/metadata.json
ADDED
|
@@ -0,0 +1,363 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"new_embodiment": {
|
| 3 |
+
"statistics": {
|
| 4 |
+
"state": {
|
| 5 |
+
"left_arm": {
|
| 6 |
+
"max": [
|
| 7 |
+
78.35006713867188,
|
| 8 |
+
73.41722106933594,
|
| 9 |
+
100.0,
|
| 10 |
+
75.20030212402344,
|
| 11 |
+
99.85347747802734
|
| 12 |
+
],
|
| 13 |
+
"min": [
|
| 14 |
+
-21.25709342956543,
|
| 15 |
+
-99.91614532470703,
|
| 16 |
+
-34.46580123901367,
|
| 17 |
+
-98.9417953491211,
|
| 18 |
+
-95.9951171875
|
| 19 |
+
],
|
| 20 |
+
"mean": [
|
| 21 |
+
22.03197479248047,
|
| 22 |
+
-36.339046478271484,
|
| 23 |
+
68.31412506103516,
|
| 24 |
+
5.61252498626709,
|
| 25 |
+
-22.224000930786133
|
| 26 |
+
],
|
| 27 |
+
"std": [
|
| 28 |
+
16.480056762695312,
|
| 29 |
+
58.42597961425781,
|
| 30 |
+
33.54707336425781,
|
| 31 |
+
54.855201721191406,
|
| 32 |
+
32.123435974121094
|
| 33 |
+
],
|
| 34 |
+
"q01": [
|
| 35 |
+
-6.453950185775753,
|
| 36 |
+
-99.24528503417969,
|
| 37 |
+
-16.04827880859375,
|
| 38 |
+
-98.9417953491211,
|
| 39 |
+
-63.19560352325439
|
| 40 |
+
],
|
| 41 |
+
"q99": [
|
| 42 |
+
78.35006713867188,
|
| 43 |
+
62.767295837402344,
|
| 44 |
+
99.63948059082031,
|
| 45 |
+
74.149658203125,
|
| 46 |
+
99.85347747802734
|
| 47 |
+
]
|
| 48 |
+
},
|
| 49 |
+
"gripper1": {
|
| 50 |
+
"max": [
|
| 51 |
+
97.7272720336914
|
| 52 |
+
],
|
| 53 |
+
"min": [
|
| 54 |
+
0.14204545319080353
|
| 55 |
+
],
|
| 56 |
+
"mean": [
|
| 57 |
+
7.044379234313965
|
| 58 |
+
],
|
| 59 |
+
"std": [
|
| 60 |
+
14.734519004821777
|
| 61 |
+
],
|
| 62 |
+
"q01": [
|
| 63 |
+
0.14204545319080353
|
| 64 |
+
],
|
| 65 |
+
"q99": [
|
| 66 |
+
74.32031150817873
|
| 67 |
+
]
|
| 68 |
+
},
|
| 69 |
+
"right_arm": {
|
| 70 |
+
"max": [
|
| 71 |
+
7.535795211791992,
|
| 72 |
+
84.6780776977539,
|
| 73 |
+
99.46284484863281,
|
| 74 |
+
74.34782409667969,
|
| 75 |
+
30.256410598754883
|
| 76 |
+
],
|
| 77 |
+
"min": [
|
| 78 |
+
-31.602048873901367,
|
| 79 |
+
-98.00664520263672,
|
| 80 |
+
-39.03837585449219,
|
| 81 |
+
-98.17805480957031,
|
| 82 |
+
-67.23494720458984
|
| 83 |
+
],
|
| 84 |
+
"mean": [
|
| 85 |
+
-3.5733110904693604,
|
| 86 |
+
-42.38075637817383,
|
| 87 |
+
75.2816390991211,
|
| 88 |
+
9.783737182617188,
|
| 89 |
+
-44.31592559814453
|
| 90 |
+
],
|
| 91 |
+
"std": [
|
| 92 |
+
10.12617015838623,
|
| 93 |
+
69.33080291748047,
|
| 94 |
+
31.52765464782715,
|
| 95 |
+
71.03961181640625,
|
| 96 |
+
25.092512130737305
|
| 97 |
+
],
|
| 98 |
+
"q01": [
|
| 99 |
+
-24.43306541442871,
|
| 100 |
+
-97.91144561767578,
|
| 101 |
+
-7.278341293334961,
|
| 102 |
+
-97.5155258178711,
|
| 103 |
+
-61.98304748535156
|
| 104 |
+
],
|
| 105 |
+
"q99": [
|
| 106 |
+
6.932931423187256,
|
| 107 |
+
68.21515655517578,
|
| 108 |
+
99.46284484863281,
|
| 109 |
+
74.34782409667969,
|
| 110 |
+
20.407204513549807
|
| 111 |
+
]
|
| 112 |
+
},
|
| 113 |
+
"gripper2": {
|
| 114 |
+
"max": [
|
| 115 |
+
71.49090576171875
|
| 116 |
+
],
|
| 117 |
+
"min": [
|
| 118 |
+
1.3722126483917236
|
| 119 |
+
],
|
| 120 |
+
"mean": [
|
| 121 |
+
12.590181350708008
|
| 122 |
+
],
|
| 123 |
+
"std": [
|
| 124 |
+
11.938528060913086
|
| 125 |
+
],
|
| 126 |
+
"q01": [
|
| 127 |
+
1.5556938648223877
|
| 128 |
+
],
|
| 129 |
+
"q99": [
|
| 130 |
+
48.94545364379883
|
| 131 |
+
]
|
| 132 |
+
}
|
| 133 |
+
},
|
| 134 |
+
"action": {
|
| 135 |
+
"left_arm": {
|
| 136 |
+
"max": [
|
| 137 |
+
78.18477630615234,
|
| 138 |
+
81.9516372680664,
|
| 139 |
+
100.0,
|
| 140 |
+
78.87323760986328,
|
| 141 |
+
100.0
|
| 142 |
+
],
|
| 143 |
+
"min": [
|
| 144 |
+
-21.44892120361328,
|
| 145 |
+
-100.0,
|
| 146 |
+
-37.03024673461914,
|
| 147 |
+
-100.0,
|
| 148 |
+
-100.0
|
| 149 |
+
],
|
| 150 |
+
"mean": [
|
| 151 |
+
22.302133560180664,
|
| 152 |
+
-36.885658264160156,
|
| 153 |
+
67.65310668945312,
|
| 154 |
+
5.681365013122559,
|
| 155 |
+
-22.29690170288086
|
| 156 |
+
],
|
| 157 |
+
"std": [
|
| 158 |
+
16.664201736450195,
|
| 159 |
+
57.817176818847656,
|
| 160 |
+
34.0733642578125,
|
| 161 |
+
55.35456466674805,
|
| 162 |
+
32.176795959472656
|
| 163 |
+
],
|
| 164 |
+
"q01": [
|
| 165 |
+
-6.389906406402588,
|
| 166 |
+
-99.91173553466797,
|
| 167 |
+
-17.40612455368042,
|
| 168 |
+
-100.0,
|
| 169 |
+
-64.58546447753906
|
| 170 |
+
],
|
| 171 |
+
"q99": [
|
| 172 |
+
78.18477630615234,
|
| 173 |
+
62.26213584899905,
|
| 174 |
+
99.81668090820312,
|
| 175 |
+
76.96768951416016,
|
| 176 |
+
100.0
|
| 177 |
+
]
|
| 178 |
+
},
|
| 179 |
+
"gripper1": {
|
| 180 |
+
"max": [
|
| 181 |
+
97.76386260986328
|
| 182 |
+
],
|
| 183 |
+
"min": [
|
| 184 |
+
0.0
|
| 185 |
+
],
|
| 186 |
+
"mean": [
|
| 187 |
+
5.872297286987305
|
| 188 |
+
],
|
| 189 |
+
"std": [
|
| 190 |
+
15.169022560119629
|
| 191 |
+
],
|
| 192 |
+
"q01": [
|
| 193 |
+
0.0
|
| 194 |
+
],
|
| 195 |
+
"q99": [
|
| 196 |
+
74.85331123352056
|
| 197 |
+
]
|
| 198 |
+
},
|
| 199 |
+
"right_arm": {
|
| 200 |
+
"max": [
|
| 201 |
+
8.018280029296875,
|
| 202 |
+
84.90252685546875,
|
| 203 |
+
99.72960662841797,
|
| 204 |
+
74.73233032226562,
|
| 205 |
+
30.77525520324707
|
| 206 |
+
],
|
| 207 |
+
"min": [
|
| 208 |
+
-31.616119384765625,
|
| 209 |
+
-99.087890625,
|
| 210 |
+
-39.26325225830078,
|
| 211 |
+
-99.57283020019531,
|
| 212 |
+
-67.18055725097656
|
| 213 |
+
],
|
| 214 |
+
"mean": [
|
| 215 |
+
-3.350882053375244,
|
| 216 |
+
-42.753501892089844,
|
| 217 |
+
75.27497100830078,
|
| 218 |
+
9.843159675598145,
|
| 219 |
+
-44.773658752441406
|
| 220 |
+
],
|
| 221 |
+
"std": [
|
| 222 |
+
10.194260597229004,
|
| 223 |
+
68.84522247314453,
|
| 224 |
+
32.172489166259766,
|
| 225 |
+
71.25251770019531,
|
| 226 |
+
25.30291748046875
|
| 227 |
+
],
|
| 228 |
+
"q01": [
|
| 229 |
+
-24.304113388061523,
|
| 230 |
+
-97.5953598022461,
|
| 231 |
+
-8.574124078750605,
|
| 232 |
+
-97.60785675048828,
|
| 233 |
+
-61.92479705810547
|
| 234 |
+
],
|
| 235 |
+
"q99": [
|
| 236 |
+
7.213930130004883,
|
| 237 |
+
69.39029693603516,
|
| 238 |
+
99.72960662841797,
|
| 239 |
+
74.73233032226562,
|
| 240 |
+
20.75176239013672
|
| 241 |
+
]
|
| 242 |
+
},
|
| 243 |
+
"gripper2": {
|
| 244 |
+
"max": [
|
| 245 |
+
72.06661987304688
|
| 246 |
+
],
|
| 247 |
+
"min": [
|
| 248 |
+
0.0
|
| 249 |
+
],
|
| 250 |
+
"mean": [
|
| 251 |
+
11.008831977844238
|
| 252 |
+
],
|
| 253 |
+
"std": [
|
| 254 |
+
12.901287078857422
|
| 255 |
+
],
|
| 256 |
+
"q01": [
|
| 257 |
+
0.9084027409553528
|
| 258 |
+
],
|
| 259 |
+
"q99": [
|
| 260 |
+
48.82664489746094
|
| 261 |
+
]
|
| 262 |
+
}
|
| 263 |
+
}
|
| 264 |
+
},
|
| 265 |
+
"modalities": {
|
| 266 |
+
"video": {
|
| 267 |
+
"right": {
|
| 268 |
+
"resolution": [
|
| 269 |
+
640,
|
| 270 |
+
480
|
| 271 |
+
],
|
| 272 |
+
"channels": 3,
|
| 273 |
+
"fps": 30.0
|
| 274 |
+
},
|
| 275 |
+
"left": {
|
| 276 |
+
"resolution": [
|
| 277 |
+
640,
|
| 278 |
+
480
|
| 279 |
+
],
|
| 280 |
+
"channels": 3,
|
| 281 |
+
"fps": 30.0
|
| 282 |
+
},
|
| 283 |
+
"top_depth": {
|
| 284 |
+
"resolution": [
|
| 285 |
+
640,
|
| 286 |
+
480
|
| 287 |
+
],
|
| 288 |
+
"channels": 3,
|
| 289 |
+
"fps": 30.0
|
| 290 |
+
}
|
| 291 |
+
},
|
| 292 |
+
"state": {
|
| 293 |
+
"left_arm": {
|
| 294 |
+
"absolute": true,
|
| 295 |
+
"rotation_type": null,
|
| 296 |
+
"shape": [
|
| 297 |
+
5
|
| 298 |
+
],
|
| 299 |
+
"continuous": true
|
| 300 |
+
},
|
| 301 |
+
"gripper1": {
|
| 302 |
+
"absolute": true,
|
| 303 |
+
"rotation_type": null,
|
| 304 |
+
"shape": [
|
| 305 |
+
1
|
| 306 |
+
],
|
| 307 |
+
"continuous": true
|
| 308 |
+
},
|
| 309 |
+
"right_arm": {
|
| 310 |
+
"absolute": true,
|
| 311 |
+
"rotation_type": null,
|
| 312 |
+
"shape": [
|
| 313 |
+
5
|
| 314 |
+
],
|
| 315 |
+
"continuous": true
|
| 316 |
+
},
|
| 317 |
+
"gripper2": {
|
| 318 |
+
"absolute": true,
|
| 319 |
+
"rotation_type": null,
|
| 320 |
+
"shape": [
|
| 321 |
+
1
|
| 322 |
+
],
|
| 323 |
+
"continuous": true
|
| 324 |
+
}
|
| 325 |
+
},
|
| 326 |
+
"action": {
|
| 327 |
+
"left_arm": {
|
| 328 |
+
"absolute": true,
|
| 329 |
+
"rotation_type": null,
|
| 330 |
+
"shape": [
|
| 331 |
+
5
|
| 332 |
+
],
|
| 333 |
+
"continuous": true
|
| 334 |
+
},
|
| 335 |
+
"gripper1": {
|
| 336 |
+
"absolute": true,
|
| 337 |
+
"rotation_type": null,
|
| 338 |
+
"shape": [
|
| 339 |
+
1
|
| 340 |
+
],
|
| 341 |
+
"continuous": true
|
| 342 |
+
},
|
| 343 |
+
"right_arm": {
|
| 344 |
+
"absolute": true,
|
| 345 |
+
"rotation_type": null,
|
| 346 |
+
"shape": [
|
| 347 |
+
5
|
| 348 |
+
],
|
| 349 |
+
"continuous": true
|
| 350 |
+
},
|
| 351 |
+
"gripper2": {
|
| 352 |
+
"absolute": true,
|
| 353 |
+
"rotation_type": null,
|
| 354 |
+
"shape": [
|
| 355 |
+
1
|
| 356 |
+
],
|
| 357 |
+
"continuous": true
|
| 358 |
+
}
|
| 359 |
+
}
|
| 360 |
+
},
|
| 361 |
+
"embodiment_tag": "new_embodiment"
|
| 362 |
+
}
|
| 363 |
+
}
|
model-00001-of-00002.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:53a9f94450b88e60e06b68c180f9f845b29739a5fdec63452a4aa6faaea0103f
|
| 3 |
+
size 4999367032
|
model-00002-of-00002.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1a9879cbdd59b4903abfed12758a5aac7e8c0a904003ef93b5200151a861cd6d
|
| 3 |
+
size 2586705312
|
model.safetensors.index.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
trainer_state.json
ADDED
|
@@ -0,0 +1,393 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": null,
|
| 3 |
+
"best_metric": null,
|
| 4 |
+
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 0.07399459839431721,
|
| 6 |
+
"eval_steps": 500,
|
| 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 |
+
"epoch": 0.0014798919678863443,
|
| 14 |
+
"grad_norm": 3.318990468978882,
|
| 15 |
+
"learning_rate": 3.6e-05,
|
| 16 |
+
"loss": 0.8225,
|
| 17 |
+
"step": 10
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.0029597839357726886,
|
| 21 |
+
"grad_norm": 1.644492745399475,
|
| 22 |
+
"learning_rate": 7.6e-05,
|
| 23 |
+
"loss": 0.2495,
|
| 24 |
+
"step": 20
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.004439675903659033,
|
| 28 |
+
"grad_norm": 2.137453556060791,
|
| 29 |
+
"learning_rate": 9.998250366089848e-05,
|
| 30 |
+
"loss": 0.2076,
|
| 31 |
+
"step": 30
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.005919567871545377,
|
| 35 |
+
"grad_norm": 1.6013213396072388,
|
| 36 |
+
"learning_rate": 9.97858104436822e-05,
|
| 37 |
+
"loss": 0.1753,
|
| 38 |
+
"step": 40
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.007399459839431722,
|
| 42 |
+
"grad_norm": 2.4884767532348633,
|
| 43 |
+
"learning_rate": 9.937141654477528e-05,
|
| 44 |
+
"loss": 0.1661,
|
| 45 |
+
"step": 50
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.008879351807318065,
|
| 49 |
+
"grad_norm": 1.5080770254135132,
|
| 50 |
+
"learning_rate": 9.87411340032603e-05,
|
| 51 |
+
"loss": 0.1414,
|
| 52 |
+
"step": 60
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.01035924377520441,
|
| 56 |
+
"grad_norm": 1.3873870372772217,
|
| 57 |
+
"learning_rate": 9.789771888432375e-05,
|
| 58 |
+
"loss": 0.1409,
|
| 59 |
+
"step": 70
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 0.011839135743090754,
|
| 63 |
+
"grad_norm": 1.1532950401306152,
|
| 64 |
+
"learning_rate": 9.684485922768422e-05,
|
| 65 |
+
"loss": 0.1169,
|
| 66 |
+
"step": 80
|
| 67 |
+
},
|
| 68 |
+
{
|
| 69 |
+
"epoch": 0.013319027710977099,
|
| 70 |
+
"grad_norm": 1.088762640953064,
|
| 71 |
+
"learning_rate": 9.558715892073323e-05,
|
| 72 |
+
"loss": 0.1043,
|
| 73 |
+
"step": 90
|
| 74 |
+
},
|
| 75 |
+
{
|
| 76 |
+
"epoch": 0.014798919678863444,
|
| 77 |
+
"grad_norm": 2.323737621307373,
|
| 78 |
+
"learning_rate": 9.413011756690685e-05,
|
| 79 |
+
"loss": 0.116,
|
| 80 |
+
"step": 100
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.016278811646749786,
|
| 84 |
+
"grad_norm": 1.2114522457122803,
|
| 85 |
+
"learning_rate": 9.248010643731935e-05,
|
| 86 |
+
"loss": 0.1146,
|
| 87 |
+
"step": 110
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"epoch": 0.01775870361463613,
|
| 91 |
+
"grad_norm": 1.2140398025512695,
|
| 92 |
+
"learning_rate": 9.064434061081562e-05,
|
| 93 |
+
"loss": 0.1248,
|
| 94 |
+
"step": 120
|
| 95 |
+
},
|
| 96 |
+
{
|
| 97 |
+
"epoch": 0.019238595582522475,
|
| 98 |
+
"grad_norm": 1.2787519693374634,
|
| 99 |
+
"learning_rate": 8.863084742426719e-05,
|
| 100 |
+
"loss": 0.0848,
|
| 101 |
+
"step": 130
|
| 102 |
+
},
|
| 103 |
+
{
|
| 104 |
+
"epoch": 0.02071848755040882,
|
| 105 |
+
"grad_norm": 0.8212881684303284,
|
| 106 |
+
"learning_rate": 8.644843137107059e-05,
|
| 107 |
+
"loss": 0.0965,
|
| 108 |
+
"step": 140
|
| 109 |
+
},
|
| 110 |
+
{
|
| 111 |
+
"epoch": 0.022198379518295164,
|
| 112 |
+
"grad_norm": 1.0979036092758179,
|
| 113 |
+
"learning_rate": 8.410663560133784e-05,
|
| 114 |
+
"loss": 0.0922,
|
| 115 |
+
"step": 150
|
| 116 |
+
},
|
| 117 |
+
{
|
| 118 |
+
"epoch": 0.02367827148618151,
|
| 119 |
+
"grad_norm": 0.9978598356246948,
|
| 120 |
+
"learning_rate": 8.161570019212921e-05,
|
| 121 |
+
"loss": 0.0916,
|
| 122 |
+
"step": 160
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.025158163454067854,
|
| 126 |
+
"grad_norm": 0.8716151714324951,
|
| 127 |
+
"learning_rate": 7.898651737020166e-05,
|
| 128 |
+
"loss": 0.067,
|
| 129 |
+
"step": 170
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.026638055421954198,
|
| 133 |
+
"grad_norm": 1.0746519565582275,
|
| 134 |
+
"learning_rate": 7.623058388307269e-05,
|
| 135 |
+
"loss": 0.0802,
|
| 136 |
+
"step": 180
|
| 137 |
+
},
|
| 138 |
+
{
|
| 139 |
+
"epoch": 0.028117947389840543,
|
| 140 |
+
"grad_norm": 1.3861006498336792,
|
| 141 |
+
"learning_rate": 7.335995072666848e-05,
|
| 142 |
+
"loss": 0.0907,
|
| 143 |
+
"step": 190
|
| 144 |
+
},
|
| 145 |
+
{
|
| 146 |
+
"epoch": 0.029597839357726887,
|
| 147 |
+
"grad_norm": 1.093328833580017,
|
| 148 |
+
"learning_rate": 7.038717044938519e-05,
|
| 149 |
+
"loss": 0.0657,
|
| 150 |
+
"step": 200
|
| 151 |
+
},
|
| 152 |
+
{
|
| 153 |
+
"epoch": 0.03107773132561323,
|
| 154 |
+
"grad_norm": 0.8141024708747864,
|
| 155 |
+
"learning_rate": 6.732524226298841e-05,
|
| 156 |
+
"loss": 0.0613,
|
| 157 |
+
"step": 210
|
| 158 |
+
},
|
| 159 |
+
{
|
| 160 |
+
"epoch": 0.03255762329349957,
|
| 161 |
+
"grad_norm": 0.9468361139297485,
|
| 162 |
+
"learning_rate": 6.418755520036775e-05,
|
| 163 |
+
"loss": 0.0829,
|
| 164 |
+
"step": 220
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.03403751526138592,
|
| 168 |
+
"grad_norm": 0.6965128183364868,
|
| 169 |
+
"learning_rate": 6.0987829568702656e-05,
|
| 170 |
+
"loss": 0.0796,
|
| 171 |
+
"step": 230
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 0.03551740722927226,
|
| 175 |
+
"grad_norm": 1.127314567565918,
|
| 176 |
+
"learning_rate": 5.7740056954050084e-05,
|
| 177 |
+
"loss": 0.0798,
|
| 178 |
+
"step": 240
|
| 179 |
+
},
|
| 180 |
+
{
|
| 181 |
+
"epoch": 0.036997299197158606,
|
| 182 |
+
"grad_norm": 1.4155848026275635,
|
| 183 |
+
"learning_rate": 5.445843903969854e-05,
|
| 184 |
+
"loss": 0.06,
|
| 185 |
+
"step": 250
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"epoch": 0.03847719116504495,
|
| 189 |
+
"grad_norm": 0.7054800987243652,
|
| 190 |
+
"learning_rate": 5.1157325505820694e-05,
|
| 191 |
+
"loss": 0.052,
|
| 192 |
+
"step": 260
|
| 193 |
+
},
|
| 194 |
+
{
|
| 195 |
+
"epoch": 0.039957083132931295,
|
| 196 |
+
"grad_norm": 0.7660062909126282,
|
| 197 |
+
"learning_rate": 4.785115128197298e-05,
|
| 198 |
+
"loss": 0.059,
|
| 199 |
+
"step": 270
|
| 200 |
+
},
|
| 201 |
+
{
|
| 202 |
+
"epoch": 0.04143697510081764,
|
| 203 |
+
"grad_norm": 0.7886701822280884,
|
| 204 |
+
"learning_rate": 4.4554373426821374e-05,
|
| 205 |
+
"loss": 0.0536,
|
| 206 |
+
"step": 280
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.042916867068703984,
|
| 210 |
+
"grad_norm": 1.4434950351715088,
|
| 211 |
+
"learning_rate": 4.1281407911102425e-05,
|
| 212 |
+
"loss": 0.0628,
|
| 213 |
+
"step": 290
|
| 214 |
+
},
|
| 215 |
+
{
|
| 216 |
+
"epoch": 0.04439675903659033,
|
| 217 |
+
"grad_norm": 0.9001641869544983,
|
| 218 |
+
"learning_rate": 3.8046566580251e-05,
|
| 219 |
+
"loss": 0.0536,
|
| 220 |
+
"step": 300
|
| 221 |
+
},
|
| 222 |
+
{
|
| 223 |
+
"epoch": 0.04587665100447667,
|
| 224 |
+
"grad_norm": 0.7539929151535034,
|
| 225 |
+
"learning_rate": 3.4863994572341843e-05,
|
| 226 |
+
"loss": 0.0673,
|
| 227 |
+
"step": 310
|
| 228 |
+
},
|
| 229 |
+
{
|
| 230 |
+
"epoch": 0.04735654297236302,
|
| 231 |
+
"grad_norm": 0.6108214855194092,
|
| 232 |
+
"learning_rate": 3.1747608464999725e-05,
|
| 233 |
+
"loss": 0.0358,
|
| 234 |
+
"step": 320
|
| 235 |
+
},
|
| 236 |
+
{
|
| 237 |
+
"epoch": 0.04883643494024936,
|
| 238 |
+
"grad_norm": 0.4837505519390106,
|
| 239 |
+
"learning_rate": 2.8711035421746367e-05,
|
| 240 |
+
"loss": 0.041,
|
| 241 |
+
"step": 330
|
| 242 |
+
},
|
| 243 |
+
{
|
| 244 |
+
"epoch": 0.05031632690813571,
|
| 245 |
+
"grad_norm": 0.8078829050064087,
|
| 246 |
+
"learning_rate": 2.5767553603881767e-05,
|
| 247 |
+
"loss": 0.0518,
|
| 248 |
+
"step": 340
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 0.05179621887602205,
|
| 252 |
+
"grad_norm": 1.5803087949752808,
|
| 253 |
+
"learning_rate": 2.29300341084631e-05,
|
| 254 |
+
"loss": 0.0486,
|
| 255 |
+
"step": 350
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 0.053276110843908396,
|
| 259 |
+
"grad_norm": 0.3863455355167389,
|
| 260 |
+
"learning_rate": 2.0210884686272368e-05,
|
| 261 |
+
"loss": 0.039,
|
| 262 |
+
"step": 360
|
| 263 |
+
},
|
| 264 |
+
{
|
| 265 |
+
"epoch": 0.05475600281179474,
|
| 266 |
+
"grad_norm": 0.7085531949996948,
|
| 267 |
+
"learning_rate": 1.7621995485879062e-05,
|
| 268 |
+
"loss": 0.0389,
|
| 269 |
+
"step": 370
|
| 270 |
+
},
|
| 271 |
+
{
|
| 272 |
+
"epoch": 0.056235894779681085,
|
| 273 |
+
"grad_norm": 0.7141103744506836,
|
| 274 |
+
"learning_rate": 1.517468706104589e-05,
|
| 275 |
+
"loss": 0.0425,
|
| 276 |
+
"step": 380
|
| 277 |
+
},
|
| 278 |
+
{
|
| 279 |
+
"epoch": 0.05771578674756743,
|
| 280 |
+
"grad_norm": 0.6287118196487427,
|
| 281 |
+
"learning_rate": 1.2879660868827508e-05,
|
| 282 |
+
"loss": 0.0457,
|
| 283 |
+
"step": 390
|
| 284 |
+
},
|
| 285 |
+
{
|
| 286 |
+
"epoch": 0.059195678715453774,
|
| 287 |
+
"grad_norm": 0.821704089641571,
|
| 288 |
+
"learning_rate": 1.0746952474821614e-05,
|
| 289 |
+
"loss": 0.0436,
|
| 290 |
+
"step": 400
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 0.06067557068334012,
|
| 294 |
+
"grad_norm": 0.6595329642295837,
|
| 295 |
+
"learning_rate": 8.785887670194138e-06,
|
| 296 |
+
"loss": 0.0436,
|
| 297 |
+
"step": 410
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 0.06215546265122646,
|
| 301 |
+
"grad_norm": 0.5961970090866089,
|
| 302 |
+
"learning_rate": 7.005041692367154e-06,
|
| 303 |
+
"loss": 0.0309,
|
| 304 |
+
"step": 420
|
| 305 |
+
},
|
| 306 |
+
{
|
| 307 |
+
"epoch": 0.06363535461911281,
|
| 308 |
+
"grad_norm": 0.578628659248352,
|
| 309 |
+
"learning_rate": 5.412201727687644e-06,
|
| 310 |
+
"loss": 0.0402,
|
| 311 |
+
"step": 430
|
| 312 |
+
},
|
| 313 |
+
{
|
| 314 |
+
"epoch": 0.06511524658699915,
|
| 315 |
+
"grad_norm": 0.48573747277259827,
|
| 316 |
+
"learning_rate": 4.01433286004283e-06,
|
| 317 |
+
"loss": 0.0338,
|
| 318 |
+
"step": 440
|
| 319 |
+
},
|
| 320 |
+
{
|
| 321 |
+
"epoch": 0.0665951385548855,
|
| 322 |
+
"grad_norm": 0.54139244556427,
|
| 323 |
+
"learning_rate": 2.817547614320615e-06,
|
| 324 |
+
"loss": 0.0269,
|
| 325 |
+
"step": 450
|
| 326 |
+
},
|
| 327 |
+
{
|
| 328 |
+
"epoch": 0.06807503052277183,
|
| 329 |
+
"grad_norm": 0.6104289293289185,
|
| 330 |
+
"learning_rate": 1.8270792278934302e-06,
|
| 331 |
+
"loss": 0.0328,
|
| 332 |
+
"step": 460
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 0.06955492249065819,
|
| 336 |
+
"grad_norm": 0.40234851837158203,
|
| 337 |
+
"learning_rate": 1.0472587670027678e-06,
|
| 338 |
+
"loss": 0.0341,
|
| 339 |
+
"step": 470
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 0.07103481445854452,
|
| 343 |
+
"grad_norm": 0.3963511288166046,
|
| 344 |
+
"learning_rate": 4.814961881085045e-07,
|
| 345 |
+
"loss": 0.0347,
|
| 346 |
+
"step": 480
|
| 347 |
+
},
|
| 348 |
+
{
|
| 349 |
+
"epoch": 0.07251470642643087,
|
| 350 |
+
"grad_norm": 0.36650583148002625,
|
| 351 |
+
"learning_rate": 1.3226542701689215e-07,
|
| 352 |
+
"loss": 0.0401,
|
| 353 |
+
"step": 490
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"epoch": 0.07399459839431721,
|
| 357 |
+
"grad_norm": 0.48519328236579895,
|
| 358 |
+
"learning_rate": 1.0935809887702154e-09,
|
| 359 |
+
"loss": 0.0292,
|
| 360 |
+
"step": 500
|
| 361 |
+
},
|
| 362 |
+
{
|
| 363 |
+
"epoch": 0.07399459839431721,
|
| 364 |
+
"step": 500,
|
| 365 |
+
"total_flos": 0.0,
|
| 366 |
+
"train_loss": 0.09187581622600556,
|
| 367 |
+
"train_runtime": 831.2312,
|
| 368 |
+
"train_samples_per_second": 4.812,
|
| 369 |
+
"train_steps_per_second": 0.602
|
| 370 |
+
}
|
| 371 |
+
],
|
| 372 |
+
"logging_steps": 10,
|
| 373 |
+
"max_steps": 500,
|
| 374 |
+
"num_input_tokens_seen": 0,
|
| 375 |
+
"num_train_epochs": 1,
|
| 376 |
+
"save_steps": 1000,
|
| 377 |
+
"stateful_callbacks": {
|
| 378 |
+
"TrainerControl": {
|
| 379 |
+
"args": {
|
| 380 |
+
"should_epoch_stop": false,
|
| 381 |
+
"should_evaluate": false,
|
| 382 |
+
"should_log": false,
|
| 383 |
+
"should_save": true,
|
| 384 |
+
"should_training_stop": true
|
| 385 |
+
},
|
| 386 |
+
"attributes": {}
|
| 387 |
+
}
|
| 388 |
+
},
|
| 389 |
+
"total_flos": 0.0,
|
| 390 |
+
"train_batch_size": 1,
|
| 391 |
+
"trial_name": null,
|
| 392 |
+
"trial_params": null
|
| 393 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6a223b435fa801ad0737013021f2f36d1666a32d0fb733b1ee86a310c949ddae
|
| 3 |
+
size 5713
|