Upload folder using huggingface_hub
Browse files- .gitattributes +1 -0
- checkpoint-1000/config.json +64 -0
- checkpoint-1000/experiment_cfg/metadata.json +259 -0
- checkpoint-1000/model-00001-of-00002.safetensors +3 -0
- checkpoint-1000/model-00002-of-00002.safetensors +3 -0
- checkpoint-1000/model.safetensors.index.json +0 -0
- checkpoint-1000/optimizer.pt +3 -0
- checkpoint-1000/rng_state.pth +3 -0
- checkpoint-1000/scheduler.pt +3 -0
- checkpoint-1000/trainer_state.json +734 -0
- config.json +64 -0
- eval_plot.png +3 -0
- experiment_cfg/metadata.json +259 -0
- model-00001-of-00002.safetensors +3 -0
- model-00002-of-00002.safetensors +3 -0
- model.safetensors.index.json +0 -0
- trainer_state.json +743 -0
- training_args.bin +3 -0
.gitattributes
CHANGED
|
@@ -33,3 +33,4 @@ 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 |
+
eval_plot.png filter=lfs diff=lfs merge=lfs -text
|
checkpoint-1000/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 |
+
}
|
checkpoint-1000/experiment_cfg/metadata.json
ADDED
|
@@ -0,0 +1,259 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"oxe_droid": {
|
| 3 |
+
"statistics": {
|
| 4 |
+
"state": {
|
| 5 |
+
"eef_position": {
|
| 6 |
+
"max": [
|
| 7 |
+
0.46507203578948975,
|
| 8 |
+
1.447200059890747,
|
| 9 |
+
0.290057897567749
|
| 10 |
+
],
|
| 11 |
+
"min": [
|
| 12 |
+
-0.40590497851371765,
|
| 13 |
+
-0.024546854197978973,
|
| 14 |
+
-0.22337864339351654
|
| 15 |
+
],
|
| 16 |
+
"mean": [
|
| 17 |
+
0.05521393194794655,
|
| 18 |
+
0.6310628056526184,
|
| 19 |
+
-0.01189372967928648
|
| 20 |
+
],
|
| 21 |
+
"std": [
|
| 22 |
+
0.18417152762413025,
|
| 23 |
+
0.2100915163755417,
|
| 24 |
+
0.0770791620016098
|
| 25 |
+
],
|
| 26 |
+
"q01": [
|
| 27 |
+
-0.37289209306240084,
|
| 28 |
+
0.14116937458515166,
|
| 29 |
+
-0.1820696112513542
|
| 30 |
+
],
|
| 31 |
+
"q99": [
|
| 32 |
+
0.39412812441587447,
|
| 33 |
+
1.09063316822052,
|
| 34 |
+
0.1658253240585327
|
| 35 |
+
]
|
| 36 |
+
},
|
| 37 |
+
"eef_rotation": {
|
| 38 |
+
"max": [
|
| 39 |
+
-0.35259532928466797,
|
| 40 |
+
0.8778502345085144,
|
| 41 |
+
2.965937376022339
|
| 42 |
+
],
|
| 43 |
+
"min": [
|
| 44 |
+
-2.3446121215820312,
|
| 45 |
+
-0.47837626934051514,
|
| 46 |
+
1.515358328819275
|
| 47 |
+
],
|
| 48 |
+
"mean": [
|
| 49 |
+
-1.6130940914154053,
|
| 50 |
+
-0.06260889768600464,
|
| 51 |
+
2.2893757820129395
|
| 52 |
+
],
|
| 53 |
+
"std": [
|
| 54 |
+
0.3057422637939453,
|
| 55 |
+
0.13727501034736633,
|
| 56 |
+
0.18679431080818176
|
| 57 |
+
],
|
| 58 |
+
"q01": [
|
| 59 |
+
-2.151788368225098,
|
| 60 |
+
-0.348803762793541,
|
| 61 |
+
1.8247082424163819
|
| 62 |
+
],
|
| 63 |
+
"q99": [
|
| 64 |
+
-0.8719659841060661,
|
| 65 |
+
0.32717462211847265,
|
| 66 |
+
2.695855288505554
|
| 67 |
+
]
|
| 68 |
+
},
|
| 69 |
+
"gripper_position": {
|
| 70 |
+
"max": [
|
| 71 |
+
1.0660839080810547
|
| 72 |
+
],
|
| 73 |
+
"min": [
|
| 74 |
+
-0.862152636051178
|
| 75 |
+
],
|
| 76 |
+
"mean": [
|
| 77 |
+
0.08788274973630905
|
| 78 |
+
],
|
| 79 |
+
"std": [
|
| 80 |
+
0.2687181830406189
|
| 81 |
+
],
|
| 82 |
+
"q01": [
|
| 83 |
+
-0.6035337620973587
|
| 84 |
+
],
|
| 85 |
+
"q99": [
|
| 86 |
+
0.7679916465282439
|
| 87 |
+
]
|
| 88 |
+
}
|
| 89 |
+
},
|
| 90 |
+
"action": {
|
| 91 |
+
"eef_position_delta": {
|
| 92 |
+
"max": [
|
| 93 |
+
0.5925613045692444,
|
| 94 |
+
0.8723994493484497,
|
| 95 |
+
0.69416743516922
|
| 96 |
+
],
|
| 97 |
+
"min": [
|
| 98 |
+
-0.960589587688446,
|
| 99 |
+
-0.8221790194511414,
|
| 100 |
+
-0.8575763702392578
|
| 101 |
+
],
|
| 102 |
+
"mean": [
|
| 103 |
+
0.03402156010270119,
|
| 104 |
+
-0.007893561385571957,
|
| 105 |
+
0.0033362761605530977
|
| 106 |
+
],
|
| 107 |
+
"std": [
|
| 108 |
+
0.19854718446731567,
|
| 109 |
+
0.15367349982261658,
|
| 110 |
+
0.14636090397834778
|
| 111 |
+
],
|
| 112 |
+
"q01": [
|
| 113 |
+
-0.47471584677696227,
|
| 114 |
+
-0.4062542155385018,
|
| 115 |
+
-0.26967047452926635
|
| 116 |
+
],
|
| 117 |
+
"q99": [
|
| 118 |
+
0.4079366648197173,
|
| 119 |
+
0.34532205790281273,
|
| 120 |
+
0.4184985920786849
|
| 121 |
+
]
|
| 122 |
+
},
|
| 123 |
+
"eef_rotation_delta": {
|
| 124 |
+
"max": [
|
| 125 |
+
0.8240683674812317,
|
| 126 |
+
0.7224147915840149,
|
| 127 |
+
0.8612414598464966
|
| 128 |
+
],
|
| 129 |
+
"min": [
|
| 130 |
+
-0.8100513815879822,
|
| 131 |
+
-0.8826454877853394,
|
| 132 |
+
-0.9472323656082153
|
| 133 |
+
],
|
| 134 |
+
"mean": [
|
| 135 |
+
0.0014922608388587832,
|
| 136 |
+
-0.029764438048005104,
|
| 137 |
+
0.006220718380063772
|
| 138 |
+
],
|
| 139 |
+
"std": [
|
| 140 |
+
0.14293360710144043,
|
| 141 |
+
0.16134193539619446,
|
| 142 |
+
0.2356378585100174
|
| 143 |
+
],
|
| 144 |
+
"q01": [
|
| 145 |
+
-0.3907526931166649,
|
| 146 |
+
-0.4691345453262329,
|
| 147 |
+
-0.5505707603693009
|
| 148 |
+
],
|
| 149 |
+
"q99": [
|
| 150 |
+
0.37015672087669316,
|
| 151 |
+
0.39511197239160506,
|
| 152 |
+
0.573674650788307
|
| 153 |
+
]
|
| 154 |
+
},
|
| 155 |
+
"gripper_position": {
|
| 156 |
+
"max": [
|
| 157 |
+
1.0
|
| 158 |
+
],
|
| 159 |
+
"min": [
|
| 160 |
+
0.0
|
| 161 |
+
],
|
| 162 |
+
"mean": [
|
| 163 |
+
0.6008301973342896
|
| 164 |
+
],
|
| 165 |
+
"std": [
|
| 166 |
+
0.3934294879436493
|
| 167 |
+
],
|
| 168 |
+
"q01": [
|
| 169 |
+
0.07158590108156204
|
| 170 |
+
],
|
| 171 |
+
"q99": [
|
| 172 |
+
1.0
|
| 173 |
+
]
|
| 174 |
+
}
|
| 175 |
+
}
|
| 176 |
+
},
|
| 177 |
+
"modalities": {
|
| 178 |
+
"video": {
|
| 179 |
+
"exterior_image_1": {
|
| 180 |
+
"resolution": [
|
| 181 |
+
320,
|
| 182 |
+
180
|
| 183 |
+
],
|
| 184 |
+
"channels": 3,
|
| 185 |
+
"fps": 15.0
|
| 186 |
+
},
|
| 187 |
+
"exterior_image_2": {
|
| 188 |
+
"resolution": [
|
| 189 |
+
320,
|
| 190 |
+
180
|
| 191 |
+
],
|
| 192 |
+
"channels": 3,
|
| 193 |
+
"fps": 15.0
|
| 194 |
+
},
|
| 195 |
+
"wrist_image": {
|
| 196 |
+
"resolution": [
|
| 197 |
+
320,
|
| 198 |
+
180
|
| 199 |
+
],
|
| 200 |
+
"channels": 3,
|
| 201 |
+
"fps": 15.0
|
| 202 |
+
}
|
| 203 |
+
},
|
| 204 |
+
"state": {
|
| 205 |
+
"eef_position": {
|
| 206 |
+
"absolute": true,
|
| 207 |
+
"rotation_type": null,
|
| 208 |
+
"shape": [
|
| 209 |
+
3
|
| 210 |
+
],
|
| 211 |
+
"continuous": true
|
| 212 |
+
},
|
| 213 |
+
"eef_rotation": {
|
| 214 |
+
"absolute": true,
|
| 215 |
+
"rotation_type": "euler_angles_rpy",
|
| 216 |
+
"shape": [
|
| 217 |
+
3
|
| 218 |
+
],
|
| 219 |
+
"continuous": true
|
| 220 |
+
},
|
| 221 |
+
"gripper_position": {
|
| 222 |
+
"absolute": true,
|
| 223 |
+
"rotation_type": null,
|
| 224 |
+
"shape": [
|
| 225 |
+
1
|
| 226 |
+
],
|
| 227 |
+
"continuous": true
|
| 228 |
+
}
|
| 229 |
+
},
|
| 230 |
+
"action": {
|
| 231 |
+
"eef_position_delta": {
|
| 232 |
+
"absolute": false,
|
| 233 |
+
"rotation_type": null,
|
| 234 |
+
"shape": [
|
| 235 |
+
3
|
| 236 |
+
],
|
| 237 |
+
"continuous": true
|
| 238 |
+
},
|
| 239 |
+
"eef_rotation_delta": {
|
| 240 |
+
"absolute": false,
|
| 241 |
+
"rotation_type": "axis_angle",
|
| 242 |
+
"shape": [
|
| 243 |
+
3
|
| 244 |
+
],
|
| 245 |
+
"continuous": true
|
| 246 |
+
},
|
| 247 |
+
"gripper_position": {
|
| 248 |
+
"absolute": false,
|
| 249 |
+
"rotation_type": null,
|
| 250 |
+
"shape": [
|
| 251 |
+
1
|
| 252 |
+
],
|
| 253 |
+
"continuous": true
|
| 254 |
+
}
|
| 255 |
+
}
|
| 256 |
+
},
|
| 257 |
+
"embodiment_tag": "oxe_droid"
|
| 258 |
+
}
|
| 259 |
+
}
|
checkpoint-1000/model-00001-of-00002.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:da5487b7f3546322ddae34bb07dccab41feae6c722565237dc73a0f96861062b
|
| 3 |
+
size 4999367032
|
checkpoint-1000/model-00002-of-00002.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c1f2def2be205cda90f1f730f18164dec31b074b0c482d18949dc4a78c32cce4
|
| 3 |
+
size 2586705312
|
checkpoint-1000/model.safetensors.index.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
checkpoint-1000/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e1db7f95cbcb58e11b8baef3025310e916ec907e6496a5d8090505b38423a179
|
| 3 |
+
size 8550720062
|
checkpoint-1000/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ac4c57336df09ab4aef5e91d6b2a5b850e4edf3cbcfc905a25d2fff14e5301d0
|
| 3 |
+
size 14244
|
checkpoint-1000/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4077036d99500a708f700f75da24d51b5300e184ad35fda49dc5a4df5596cca2
|
| 3 |
+
size 1064
|
checkpoint-1000/trainer_state.json
ADDED
|
@@ -0,0 +1,734 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": null,
|
| 3 |
+
"best_metric": null,
|
| 4 |
+
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 1.9267822736030829,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 1000,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.019267822736030827,
|
| 14 |
+
"grad_norm": 1.7207869291305542,
|
| 15 |
+
"learning_rate": 1.8e-05,
|
| 16 |
+
"loss": 0.2916,
|
| 17 |
+
"step": 10
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.038535645472061654,
|
| 21 |
+
"grad_norm": 0.542586088180542,
|
| 22 |
+
"learning_rate": 3.8e-05,
|
| 23 |
+
"loss": 0.1243,
|
| 24 |
+
"step": 20
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.057803468208092484,
|
| 28 |
+
"grad_norm": 0.7696737051010132,
|
| 29 |
+
"learning_rate": 5.8e-05,
|
| 30 |
+
"loss": 0.1138,
|
| 31 |
+
"step": 30
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.07707129094412331,
|
| 35 |
+
"grad_norm": 0.524152398109436,
|
| 36 |
+
"learning_rate": 7.800000000000001e-05,
|
| 37 |
+
"loss": 0.1041,
|
| 38 |
+
"step": 40
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.09633911368015415,
|
| 42 |
+
"grad_norm": 0.4883207380771637,
|
| 43 |
+
"learning_rate": 9.8e-05,
|
| 44 |
+
"loss": 0.0938,
|
| 45 |
+
"step": 50
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.11560693641618497,
|
| 49 |
+
"grad_norm": 0.6807201504707336,
|
| 50 |
+
"learning_rate": 9.997785653888835e-05,
|
| 51 |
+
"loss": 0.1036,
|
| 52 |
+
"step": 60
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.1348747591522158,
|
| 56 |
+
"grad_norm": 0.5879040360450745,
|
| 57 |
+
"learning_rate": 9.990133642141359e-05,
|
| 58 |
+
"loss": 0.0913,
|
| 59 |
+
"step": 70
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 0.15414258188824662,
|
| 63 |
+
"grad_norm": 0.2884378135204315,
|
| 64 |
+
"learning_rate": 9.977024992520602e-05,
|
| 65 |
+
"loss": 0.0937,
|
| 66 |
+
"step": 80
|
| 67 |
+
},
|
| 68 |
+
{
|
| 69 |
+
"epoch": 0.17341040462427745,
|
| 70 |
+
"grad_norm": 0.31237179040908813,
|
| 71 |
+
"learning_rate": 9.95847403914247e-05,
|
| 72 |
+
"loss": 0.0818,
|
| 73 |
+
"step": 90
|
| 74 |
+
},
|
| 75 |
+
{
|
| 76 |
+
"epoch": 0.1926782273603083,
|
| 77 |
+
"grad_norm": 0.5337116718292236,
|
| 78 |
+
"learning_rate": 9.934501067202117e-05,
|
| 79 |
+
"loss": 0.0871,
|
| 80 |
+
"step": 100
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.2119460500963391,
|
| 84 |
+
"grad_norm": 0.35781776905059814,
|
| 85 |
+
"learning_rate": 9.905132290792394e-05,
|
| 86 |
+
"loss": 0.0813,
|
| 87 |
+
"step": 110
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"epoch": 0.23121387283236994,
|
| 91 |
+
"grad_norm": 0.41506218910217285,
|
| 92 |
+
"learning_rate": 9.870399824239117e-05,
|
| 93 |
+
"loss": 0.0807,
|
| 94 |
+
"step": 120
|
| 95 |
+
},
|
| 96 |
+
{
|
| 97 |
+
"epoch": 0.2504816955684008,
|
| 98 |
+
"grad_norm": 0.2967628836631775,
|
| 99 |
+
"learning_rate": 9.830341646984521e-05,
|
| 100 |
+
"loss": 0.0824,
|
| 101 |
+
"step": 130
|
| 102 |
+
},
|
| 103 |
+
{
|
| 104 |
+
"epoch": 0.2697495183044316,
|
| 105 |
+
"grad_norm": 0.31782740354537964,
|
| 106 |
+
"learning_rate": 9.785001562057309e-05,
|
| 107 |
+
"loss": 0.0756,
|
| 108 |
+
"step": 140
|
| 109 |
+
},
|
| 110 |
+
{
|
| 111 |
+
"epoch": 0.28901734104046245,
|
| 112 |
+
"grad_norm": 0.34691065549850464,
|
| 113 |
+
"learning_rate": 9.734429148174675e-05,
|
| 114 |
+
"loss": 0.084,
|
| 115 |
+
"step": 150
|
| 116 |
+
},
|
| 117 |
+
{
|
| 118 |
+
"epoch": 0.30828516377649323,
|
| 119 |
+
"grad_norm": 0.48853012919425964,
|
| 120 |
+
"learning_rate": 9.6786797055287e-05,
|
| 121 |
+
"loss": 0.0832,
|
| 122 |
+
"step": 160
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.32755298651252407,
|
| 126 |
+
"grad_norm": 0.39761826395988464,
|
| 127 |
+
"learning_rate": 9.617814195316411e-05,
|
| 128 |
+
"loss": 0.0773,
|
| 129 |
+
"step": 170
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.3468208092485549,
|
| 133 |
+
"grad_norm": 0.41890355944633484,
|
| 134 |
+
"learning_rate": 9.551899173079607e-05,
|
| 135 |
+
"loss": 0.072,
|
| 136 |
+
"step": 180
|
| 137 |
+
},
|
| 138 |
+
{
|
| 139 |
+
"epoch": 0.36608863198458574,
|
| 140 |
+
"grad_norm": 0.5497244596481323,
|
| 141 |
+
"learning_rate": 9.481006715927351e-05,
|
| 142 |
+
"loss": 0.0733,
|
| 143 |
+
"step": 190
|
| 144 |
+
},
|
| 145 |
+
{
|
| 146 |
+
"epoch": 0.3853564547206166,
|
| 147 |
+
"grad_norm": 0.28932520747184753,
|
| 148 |
+
"learning_rate": 9.405214343720707e-05,
|
| 149 |
+
"loss": 0.0814,
|
| 150 |
+
"step": 200
|
| 151 |
+
},
|
| 152 |
+
{
|
| 153 |
+
"epoch": 0.4046242774566474,
|
| 154 |
+
"grad_norm": 0.5443249344825745,
|
| 155 |
+
"learning_rate": 9.32460493430591e-05,
|
| 156 |
+
"loss": 0.0789,
|
| 157 |
+
"step": 210
|
| 158 |
+
},
|
| 159 |
+
{
|
| 160 |
+
"epoch": 0.4238921001926782,
|
| 161 |
+
"grad_norm": 0.5669806003570557,
|
| 162 |
+
"learning_rate": 9.239266632888659e-05,
|
| 163 |
+
"loss": 0.0717,
|
| 164 |
+
"step": 220
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.44315992292870904,
|
| 168 |
+
"grad_norm": 0.3382496237754822,
|
| 169 |
+
"learning_rate": 9.14929275564863e-05,
|
| 170 |
+
"loss": 0.0714,
|
| 171 |
+
"step": 230
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 0.4624277456647399,
|
| 175 |
+
"grad_norm": 0.47498270869255066,
|
| 176 |
+
"learning_rate": 9.0547816876996e-05,
|
| 177 |
+
"loss": 0.0691,
|
| 178 |
+
"step": 240
|
| 179 |
+
},
|
| 180 |
+
{
|
| 181 |
+
"epoch": 0.4816955684007707,
|
| 182 |
+
"grad_norm": 0.4482809007167816,
|
| 183 |
+
"learning_rate": 8.955836775506776e-05,
|
| 184 |
+
"loss": 0.0694,
|
| 185 |
+
"step": 250
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"epoch": 0.5009633911368016,
|
| 189 |
+
"grad_norm": 0.5501825213432312,
|
| 190 |
+
"learning_rate": 8.852566213878947e-05,
|
| 191 |
+
"loss": 0.0686,
|
| 192 |
+
"step": 260
|
| 193 |
+
},
|
| 194 |
+
{
|
| 195 |
+
"epoch": 0.5202312138728323,
|
| 196 |
+
"grad_norm": 0.3987138867378235,
|
| 197 |
+
"learning_rate": 8.745082927659047e-05,
|
| 198 |
+
"loss": 0.078,
|
| 199 |
+
"step": 270
|
| 200 |
+
},
|
| 201 |
+
{
|
| 202 |
+
"epoch": 0.5394990366088632,
|
| 203 |
+
"grad_norm": 0.28167715668678284,
|
| 204 |
+
"learning_rate": 8.633504448242505e-05,
|
| 205 |
+
"loss": 0.0758,
|
| 206 |
+
"step": 280
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.558766859344894,
|
| 210 |
+
"grad_norm": 0.24536319077014923,
|
| 211 |
+
"learning_rate": 8.517952785058385e-05,
|
| 212 |
+
"loss": 0.0747,
|
| 213 |
+
"step": 290
|
| 214 |
+
},
|
| 215 |
+
{
|
| 216 |
+
"epoch": 0.5780346820809249,
|
| 217 |
+
"grad_norm": 0.2790259122848511,
|
| 218 |
+
"learning_rate": 8.398554292153866e-05,
|
| 219 |
+
"loss": 0.0681,
|
| 220 |
+
"step": 300
|
| 221 |
+
},
|
| 222 |
+
{
|
| 223 |
+
"epoch": 0.5973025048169557,
|
| 224 |
+
"grad_norm": 0.4405696988105774,
|
| 225 |
+
"learning_rate": 8.275439530027948e-05,
|
| 226 |
+
"loss": 0.0713,
|
| 227 |
+
"step": 310
|
| 228 |
+
},
|
| 229 |
+
{
|
| 230 |
+
"epoch": 0.6165703275529865,
|
| 231 |
+
"grad_norm": 0.3930000066757202,
|
| 232 |
+
"learning_rate": 8.148743122865463e-05,
|
| 233 |
+
"loss": 0.0633,
|
| 234 |
+
"step": 320
|
| 235 |
+
},
|
| 236 |
+
{
|
| 237 |
+
"epoch": 0.6358381502890174,
|
| 238 |
+
"grad_norm": 0.39871180057525635,
|
| 239 |
+
"learning_rate": 8.018603611327504e-05,
|
| 240 |
+
"loss": 0.0735,
|
| 241 |
+
"step": 330
|
| 242 |
+
},
|
| 243 |
+
{
|
| 244 |
+
"epoch": 0.6551059730250481,
|
| 245 |
+
"grad_norm": 0.43595343828201294,
|
| 246 |
+
"learning_rate": 7.88516330105925e-05,
|
| 247 |
+
"loss": 0.0651,
|
| 248 |
+
"step": 340
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 0.674373795761079,
|
| 252 |
+
"grad_norm": 0.3049931824207306,
|
| 253 |
+
"learning_rate": 7.748568107080832e-05,
|
| 254 |
+
"loss": 0.0641,
|
| 255 |
+
"step": 350
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 0.6936416184971098,
|
| 259 |
+
"grad_norm": 0.2661554515361786,
|
| 260 |
+
"learning_rate": 7.608967394231387e-05,
|
| 261 |
+
"loss": 0.0615,
|
| 262 |
+
"step": 360
|
| 263 |
+
},
|
| 264 |
+
{
|
| 265 |
+
"epoch": 0.7129094412331407,
|
| 266 |
+
"grad_norm": 0.23182271420955658,
|
| 267 |
+
"learning_rate": 7.466513813840825e-05,
|
| 268 |
+
"loss": 0.0641,
|
| 269 |
+
"step": 370
|
| 270 |
+
},
|
| 271 |
+
{
|
| 272 |
+
"epoch": 0.7321772639691715,
|
| 273 |
+
"grad_norm": 0.36455002427101135,
|
| 274 |
+
"learning_rate": 7.32136313680782e-05,
|
| 275 |
+
"loss": 0.0694,
|
| 276 |
+
"step": 380
|
| 277 |
+
},
|
| 278 |
+
{
|
| 279 |
+
"epoch": 0.7514450867052023,
|
| 280 |
+
"grad_norm": 0.37729915976524353,
|
| 281 |
+
"learning_rate": 7.173674083266624e-05,
|
| 282 |
+
"loss": 0.0657,
|
| 283 |
+
"step": 390
|
| 284 |
+
},
|
| 285 |
+
{
|
| 286 |
+
"epoch": 0.7707129094412332,
|
| 287 |
+
"grad_norm": 0.3024880290031433,
|
| 288 |
+
"learning_rate": 7.023608149028937e-05,
|
| 289 |
+
"loss": 0.0595,
|
| 290 |
+
"step": 400
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 0.789980732177264,
|
| 294 |
+
"grad_norm": 0.31104758381843567,
|
| 295 |
+
"learning_rate": 6.871329428990602e-05,
|
| 296 |
+
"loss": 0.0681,
|
| 297 |
+
"step": 410
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 0.8092485549132948,
|
| 301 |
+
"grad_norm": 0.350875586271286,
|
| 302 |
+
"learning_rate": 6.71700443769625e-05,
|
| 303 |
+
"loss": 0.061,
|
| 304 |
+
"step": 420
|
| 305 |
+
},
|
| 306 |
+
{
|
| 307 |
+
"epoch": 0.8285163776493256,
|
| 308 |
+
"grad_norm": 0.2729007303714752,
|
| 309 |
+
"learning_rate": 6.56080192725808e-05,
|
| 310 |
+
"loss": 0.0666,
|
| 311 |
+
"step": 430
|
| 312 |
+
},
|
| 313 |
+
{
|
| 314 |
+
"epoch": 0.8477842003853564,
|
| 315 |
+
"grad_norm": 0.29519274830818176,
|
| 316 |
+
"learning_rate": 6.402892702827916e-05,
|
| 317 |
+
"loss": 0.0617,
|
| 318 |
+
"step": 440
|
| 319 |
+
},
|
| 320 |
+
{
|
| 321 |
+
"epoch": 0.8670520231213873,
|
| 322 |
+
"grad_norm": 0.41105949878692627,
|
| 323 |
+
"learning_rate": 6.243449435824276e-05,
|
| 324 |
+
"loss": 0.0635,
|
| 325 |
+
"step": 450
|
| 326 |
+
},
|
| 327 |
+
{
|
| 328 |
+
"epoch": 0.8863198458574181,
|
| 329 |
+
"grad_norm": 0.39932650327682495,
|
| 330 |
+
"learning_rate": 6.0826464751186994e-05,
|
| 331 |
+
"loss": 0.0668,
|
| 332 |
+
"step": 460
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 0.905587668593449,
|
| 336 |
+
"grad_norm": 0.36617910861968994,
|
| 337 |
+
"learning_rate": 5.9206596563878357e-05,
|
| 338 |
+
"loss": 0.0642,
|
| 339 |
+
"step": 470
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 0.9248554913294798,
|
| 343 |
+
"grad_norm": 0.3059275150299072,
|
| 344 |
+
"learning_rate": 5.757666109839702e-05,
|
| 345 |
+
"loss": 0.0639,
|
| 346 |
+
"step": 480
|
| 347 |
+
},
|
| 348 |
+
{
|
| 349 |
+
"epoch": 0.9441233140655106,
|
| 350 |
+
"grad_norm": 0.3001484274864197,
|
| 351 |
+
"learning_rate": 5.5938440665244006e-05,
|
| 352 |
+
"loss": 0.0581,
|
| 353 |
+
"step": 490
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"epoch": 0.9633911368015414,
|
| 357 |
+
"grad_norm": 0.3571004569530487,
|
| 358 |
+
"learning_rate": 5.4293726634410855e-05,
|
| 359 |
+
"loss": 0.0558,
|
| 360 |
+
"step": 500
|
| 361 |
+
},
|
| 362 |
+
{
|
| 363 |
+
"epoch": 0.9826589595375722,
|
| 364 |
+
"grad_norm": 0.3813907206058502,
|
| 365 |
+
"learning_rate": 5.264431747654284e-05,
|
| 366 |
+
"loss": 0.0618,
|
| 367 |
+
"step": 510
|
| 368 |
+
},
|
| 369 |
+
{
|
| 370 |
+
"epoch": 1.001926782273603,
|
| 371 |
+
"grad_norm": 0.29777514934539795,
|
| 372 |
+
"learning_rate": 5.0992016796337686e-05,
|
| 373 |
+
"loss": 0.0578,
|
| 374 |
+
"step": 520
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 1.0211946050096339,
|
| 378 |
+
"grad_norm": 0.28990840911865234,
|
| 379 |
+
"learning_rate": 4.93386313603304e-05,
|
| 380 |
+
"loss": 0.0613,
|
| 381 |
+
"step": 530
|
| 382 |
+
},
|
| 383 |
+
{
|
| 384 |
+
"epoch": 1.0404624277456647,
|
| 385 |
+
"grad_norm": 0.43356871604919434,
|
| 386 |
+
"learning_rate": 4.7685969121220456e-05,
|
| 387 |
+
"loss": 0.057,
|
| 388 |
+
"step": 540
|
| 389 |
+
},
|
| 390 |
+
{
|
| 391 |
+
"epoch": 1.0597302504816957,
|
| 392 |
+
"grad_norm": 0.2468729466199875,
|
| 393 |
+
"learning_rate": 4.60358372409022e-05,
|
| 394 |
+
"loss": 0.053,
|
| 395 |
+
"step": 550
|
| 396 |
+
},
|
| 397 |
+
{
|
| 398 |
+
"epoch": 1.0789980732177264,
|
| 399 |
+
"grad_norm": 0.33707353472709656,
|
| 400 |
+
"learning_rate": 4.439004011435979e-05,
|
| 401 |
+
"loss": 0.0523,
|
| 402 |
+
"step": 560
|
| 403 |
+
},
|
| 404 |
+
{
|
| 405 |
+
"epoch": 1.0982658959537572,
|
| 406 |
+
"grad_norm": 0.3425613343715668,
|
| 407 |
+
"learning_rate": 4.275037739658771e-05,
|
| 408 |
+
"loss": 0.0533,
|
| 409 |
+
"step": 570
|
| 410 |
+
},
|
| 411 |
+
{
|
| 412 |
+
"epoch": 1.117533718689788,
|
| 413 |
+
"grad_norm": 0.21212391555309296,
|
| 414 |
+
"learning_rate": 4.111864203469457e-05,
|
| 415 |
+
"loss": 0.0506,
|
| 416 |
+
"step": 580
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 1.1368015414258188,
|
| 420 |
+
"grad_norm": 0.2778816223144531,
|
| 421 |
+
"learning_rate": 3.949661830734172e-05,
|
| 422 |
+
"loss": 0.0575,
|
| 423 |
+
"step": 590
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"epoch": 1.1560693641618498,
|
| 427 |
+
"grad_norm": 0.2917283773422241,
|
| 428 |
+
"learning_rate": 3.788607987366069e-05,
|
| 429 |
+
"loss": 0.0571,
|
| 430 |
+
"step": 600
|
| 431 |
+
},
|
| 432 |
+
{
|
| 433 |
+
"epoch": 1.1753371868978806,
|
| 434 |
+
"grad_norm": 0.27486225962638855,
|
| 435 |
+
"learning_rate": 3.628878783378302e-05,
|
| 436 |
+
"loss": 0.056,
|
| 437 |
+
"step": 610
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 1.1946050096339114,
|
| 441 |
+
"grad_norm": 0.40446701645851135,
|
| 442 |
+
"learning_rate": 3.470648880310313e-05,
|
| 443 |
+
"loss": 0.0529,
|
| 444 |
+
"step": 620
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 1.2138728323699421,
|
| 448 |
+
"grad_norm": 0.25892701745033264,
|
| 449 |
+
"learning_rate": 3.3140913002379995e-05,
|
| 450 |
+
"loss": 0.0504,
|
| 451 |
+
"step": 630
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 1.2331406551059731,
|
| 455 |
+
"grad_norm": 0.26697343587875366,
|
| 456 |
+
"learning_rate": 3.1593772365766105e-05,
|
| 457 |
+
"loss": 0.0509,
|
| 458 |
+
"step": 640
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 1.252408477842004,
|
| 462 |
+
"grad_norm": 0.2691023051738739,
|
| 463 |
+
"learning_rate": 3.006675866883275e-05,
|
| 464 |
+
"loss": 0.0576,
|
| 465 |
+
"step": 650
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 1.2716763005780347,
|
| 469 |
+
"grad_norm": 0.2726045250892639,
|
| 470 |
+
"learning_rate": 2.8561541678638142e-05,
|
| 471 |
+
"loss": 0.051,
|
| 472 |
+
"step": 660
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 1.2909441233140655,
|
| 476 |
+
"grad_norm": 0.3852327764034271,
|
| 477 |
+
"learning_rate": 2.707976732786166e-05,
|
| 478 |
+
"loss": 0.0535,
|
| 479 |
+
"step": 670
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 1.3102119460500963,
|
| 483 |
+
"grad_norm": 0.2920765280723572,
|
| 484 |
+
"learning_rate": 2.562305591500069e-05,
|
| 485 |
+
"loss": 0.0492,
|
| 486 |
+
"step": 680
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 1.3294797687861273,
|
| 490 |
+
"grad_norm": 0.1913856863975525,
|
| 491 |
+
"learning_rate": 2.419300033259798e-05,
|
| 492 |
+
"loss": 0.0507,
|
| 493 |
+
"step": 690
|
| 494 |
+
},
|
| 495 |
+
{
|
| 496 |
+
"epoch": 1.348747591522158,
|
| 497 |
+
"grad_norm": 0.29561716318130493,
|
| 498 |
+
"learning_rate": 2.279116432543705e-05,
|
| 499 |
+
"loss": 0.0499,
|
| 500 |
+
"step": 700
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 1.3680154142581888,
|
| 504 |
+
"grad_norm": 0.4290018379688263,
|
| 505 |
+
"learning_rate": 2.1419080780610123e-05,
|
| 506 |
+
"loss": 0.0571,
|
| 507 |
+
"step": 710
|
| 508 |
+
},
|
| 509 |
+
{
|
| 510 |
+
"epoch": 1.3872832369942196,
|
| 511 |
+
"grad_norm": 0.2543991506099701,
|
| 512 |
+
"learning_rate": 2.0078250051328784e-05,
|
| 513 |
+
"loss": 0.0524,
|
| 514 |
+
"step": 720
|
| 515 |
+
},
|
| 516 |
+
{
|
| 517 |
+
"epoch": 1.4065510597302504,
|
| 518 |
+
"grad_norm": 0.2487826943397522,
|
| 519 |
+
"learning_rate": 1.877013831630961e-05,
|
| 520 |
+
"loss": 0.0539,
|
| 521 |
+
"step": 730
|
| 522 |
+
},
|
| 523 |
+
{
|
| 524 |
+
"epoch": 1.4258188824662814,
|
| 525 |
+
"grad_norm": 0.3550814390182495,
|
| 526 |
+
"learning_rate": 1.749617597652934e-05,
|
| 527 |
+
"loss": 0.0436,
|
| 528 |
+
"step": 740
|
| 529 |
+
},
|
| 530 |
+
{
|
| 531 |
+
"epoch": 1.4450867052023122,
|
| 532 |
+
"grad_norm": 0.2873481810092926,
|
| 533 |
+
"learning_rate": 1.62577560911024e-05,
|
| 534 |
+
"loss": 0.0463,
|
| 535 |
+
"step": 750
|
| 536 |
+
},
|
| 537 |
+
{
|
| 538 |
+
"epoch": 1.464354527938343,
|
| 539 |
+
"grad_norm": 0.22322338819503784,
|
| 540 |
+
"learning_rate": 1.5056232853991209e-05,
|
| 541 |
+
"loss": 0.0427,
|
| 542 |
+
"step": 760
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 1.4836223506743738,
|
| 546 |
+
"grad_norm": 0.23233197629451752,
|
| 547 |
+
"learning_rate": 1.389292011321498e-05,
|
| 548 |
+
"loss": 0.0427,
|
| 549 |
+
"step": 770
|
| 550 |
+
},
|
| 551 |
+
{
|
| 552 |
+
"epoch": 1.5028901734104045,
|
| 553 |
+
"grad_norm": 0.25211092829704285,
|
| 554 |
+
"learning_rate": 1.2769089934176126e-05,
|
| 555 |
+
"loss": 0.0456,
|
| 556 |
+
"step": 780
|
| 557 |
+
},
|
| 558 |
+
{
|
| 559 |
+
"epoch": 1.5221579961464355,
|
| 560 |
+
"grad_norm": 0.19579973816871643,
|
| 561 |
+
"learning_rate": 1.1685971208675539e-05,
|
| 562 |
+
"loss": 0.0447,
|
| 563 |
+
"step": 790
|
| 564 |
+
},
|
| 565 |
+
{
|
| 566 |
+
"epoch": 1.5414258188824663,
|
| 567 |
+
"grad_norm": 0.2609504461288452,
|
| 568 |
+
"learning_rate": 1.0644748311137376e-05,
|
| 569 |
+
"loss": 0.0467,
|
| 570 |
+
"step": 800
|
| 571 |
+
},
|
| 572 |
+
{
|
| 573 |
+
"epoch": 1.560693641618497,
|
| 574 |
+
"grad_norm": 0.23123040795326233,
|
| 575 |
+
"learning_rate": 9.646559803512994e-06,
|
| 576 |
+
"loss": 0.0505,
|
| 577 |
+
"step": 810
|
| 578 |
+
},
|
| 579 |
+
{
|
| 580 |
+
"epoch": 1.579961464354528,
|
| 581 |
+
"grad_norm": 0.1902703195810318,
|
| 582 |
+
"learning_rate": 8.692497190280224e-06,
|
| 583 |
+
"loss": 0.0463,
|
| 584 |
+
"step": 820
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 1.5992292870905587,
|
| 588 |
+
"grad_norm": 0.30633774399757385,
|
| 589 |
+
"learning_rate": 7.783603724899257e-06,
|
| 590 |
+
"loss": 0.0465,
|
| 591 |
+
"step": 830
|
| 592 |
+
},
|
| 593 |
+
{
|
| 594 |
+
"epoch": 1.6184971098265897,
|
| 595 |
+
"grad_norm": 0.24036064743995667,
|
| 596 |
+
"learning_rate": 6.92087326903022e-06,
|
| 597 |
+
"loss": 0.0435,
|
| 598 |
+
"step": 840
|
| 599 |
+
},
|
| 600 |
+
{
|
| 601 |
+
"epoch": 1.6377649325626205,
|
| 602 |
+
"grad_norm": 0.20757320523262024,
|
| 603 |
+
"learning_rate": 6.1052492057601275e-06,
|
| 604 |
+
"loss": 0.0505,
|
| 605 |
+
"step": 850
|
| 606 |
+
},
|
| 607 |
+
{
|
| 608 |
+
"epoch": 1.6570327552986512,
|
| 609 |
+
"grad_norm": 0.3485427796840668,
|
| 610 |
+
"learning_rate": 5.337623408027293e-06,
|
| 611 |
+
"loss": 0.0441,
|
| 612 |
+
"step": 860
|
| 613 |
+
},
|
| 614 |
+
{
|
| 615 |
+
"epoch": 1.6763005780346822,
|
| 616 |
+
"grad_norm": 0.235531747341156,
|
| 617 |
+
"learning_rate": 4.618835263371396e-06,
|
| 618 |
+
"loss": 0.0464,
|
| 619 |
+
"step": 870
|
| 620 |
+
},
|
| 621 |
+
{
|
| 622 |
+
"epoch": 1.6955684007707128,
|
| 623 |
+
"grad_norm": 0.19758078455924988,
|
| 624 |
+
"learning_rate": 3.949670756075447e-06,
|
| 625 |
+
"loss": 0.0413,
|
| 626 |
+
"step": 880
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 1.7148362235067438,
|
| 630 |
+
"grad_norm": 0.21413107216358185,
|
| 631 |
+
"learning_rate": 3.3308616077036115e-06,
|
| 632 |
+
"loss": 0.0412,
|
| 633 |
+
"step": 890
|
| 634 |
+
},
|
| 635 |
+
{
|
| 636 |
+
"epoch": 1.7341040462427746,
|
| 637 |
+
"grad_norm": 0.26742419600486755,
|
| 638 |
+
"learning_rate": 2.7630844769743757e-06,
|
| 639 |
+
"loss": 0.0454,
|
| 640 |
+
"step": 900
|
| 641 |
+
},
|
| 642 |
+
{
|
| 643 |
+
"epoch": 1.7533718689788054,
|
| 644 |
+
"grad_norm": 0.22228221595287323,
|
| 645 |
+
"learning_rate": 2.2469602198441573e-06,
|
| 646 |
+
"loss": 0.0392,
|
| 647 |
+
"step": 910
|
| 648 |
+
},
|
| 649 |
+
{
|
| 650 |
+
"epoch": 1.7726396917148364,
|
| 651 |
+
"grad_norm": 0.2834679186344147,
|
| 652 |
+
"learning_rate": 1.7830532106104747e-06,
|
| 653 |
+
"loss": 0.0436,
|
| 654 |
+
"step": 920
|
| 655 |
+
},
|
| 656 |
+
{
|
| 657 |
+
"epoch": 1.791907514450867,
|
| 658 |
+
"grad_norm": 0.260470449924469,
|
| 659 |
+
"learning_rate": 1.3718707247769135e-06,
|
| 660 |
+
"loss": 0.0424,
|
| 661 |
+
"step": 930
|
| 662 |
+
},
|
| 663 |
+
{
|
| 664 |
+
"epoch": 1.811175337186898,
|
| 665 |
+
"grad_norm": 0.2204512655735016,
|
| 666 |
+
"learning_rate": 1.0138623843548078e-06,
|
| 667 |
+
"loss": 0.0393,
|
| 668 |
+
"step": 940
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 1.8304431599229287,
|
| 672 |
+
"grad_norm": 0.23786160349845886,
|
| 673 |
+
"learning_rate": 7.094196662081831e-07,
|
| 674 |
+
"loss": 0.0472,
|
| 675 |
+
"step": 950
|
| 676 |
+
},
|
| 677 |
+
{
|
| 678 |
+
"epoch": 1.8497109826589595,
|
| 679 |
+
"grad_norm": 0.18833257257938385,
|
| 680 |
+
"learning_rate": 4.5887547397955864e-07,
|
| 681 |
+
"loss": 0.0416,
|
| 682 |
+
"step": 960
|
| 683 |
+
},
|
| 684 |
+
{
|
| 685 |
+
"epoch": 1.8689788053949905,
|
| 686 |
+
"grad_norm": 0.20266398787498474,
|
| 687 |
+
"learning_rate": 2.625037740646763e-07,
|
| 688 |
+
"loss": 0.0426,
|
| 689 |
+
"step": 970
|
| 690 |
+
},
|
| 691 |
+
{
|
| 692 |
+
"epoch": 1.888246628131021,
|
| 693 |
+
"grad_norm": 0.2135821133852005,
|
| 694 |
+
"learning_rate": 1.2051929603428825e-07,
|
| 695 |
+
"loss": 0.0445,
|
| 696 |
+
"step": 980
|
| 697 |
+
},
|
| 698 |
+
{
|
| 699 |
+
"epoch": 1.907514450867052,
|
| 700 |
+
"grad_norm": 0.2030361145734787,
|
| 701 |
+
"learning_rate": 3.3077297830541584e-08,
|
| 702 |
+
"loss": 0.045,
|
| 703 |
+
"step": 990
|
| 704 |
+
},
|
| 705 |
+
{
|
| 706 |
+
"epoch": 1.9267822736030829,
|
| 707 |
+
"grad_norm": 0.31092095375061035,
|
| 708 |
+
"learning_rate": 2.7339599464326627e-10,
|
| 709 |
+
"loss": 0.0434,
|
| 710 |
+
"step": 1000
|
| 711 |
+
}
|
| 712 |
+
],
|
| 713 |
+
"logging_steps": 10,
|
| 714 |
+
"max_steps": 1000,
|
| 715 |
+
"num_input_tokens_seen": 0,
|
| 716 |
+
"num_train_epochs": 2,
|
| 717 |
+
"save_steps": 1000,
|
| 718 |
+
"stateful_callbacks": {
|
| 719 |
+
"TrainerControl": {
|
| 720 |
+
"args": {
|
| 721 |
+
"should_epoch_stop": false,
|
| 722 |
+
"should_evaluate": false,
|
| 723 |
+
"should_log": false,
|
| 724 |
+
"should_save": true,
|
| 725 |
+
"should_training_stop": true
|
| 726 |
+
},
|
| 727 |
+
"attributes": {}
|
| 728 |
+
}
|
| 729 |
+
},
|
| 730 |
+
"total_flos": 0.0,
|
| 731 |
+
"train_batch_size": 32,
|
| 732 |
+
"trial_name": null,
|
| 733 |
+
"trial_params": null
|
| 734 |
+
}
|
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 |
+
}
|
eval_plot.png
ADDED
|
Git LFS Details
|
experiment_cfg/metadata.json
ADDED
|
@@ -0,0 +1,259 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"oxe_droid": {
|
| 3 |
+
"statistics": {
|
| 4 |
+
"state": {
|
| 5 |
+
"eef_position": {
|
| 6 |
+
"max": [
|
| 7 |
+
0.46507203578948975,
|
| 8 |
+
1.447200059890747,
|
| 9 |
+
0.290057897567749
|
| 10 |
+
],
|
| 11 |
+
"min": [
|
| 12 |
+
-0.40590497851371765,
|
| 13 |
+
-0.024546854197978973,
|
| 14 |
+
-0.22337864339351654
|
| 15 |
+
],
|
| 16 |
+
"mean": [
|
| 17 |
+
0.05521393194794655,
|
| 18 |
+
0.6310628056526184,
|
| 19 |
+
-0.01189372967928648
|
| 20 |
+
],
|
| 21 |
+
"std": [
|
| 22 |
+
0.18417152762413025,
|
| 23 |
+
0.2100915163755417,
|
| 24 |
+
0.0770791620016098
|
| 25 |
+
],
|
| 26 |
+
"q01": [
|
| 27 |
+
-0.37289209306240084,
|
| 28 |
+
0.14116937458515166,
|
| 29 |
+
-0.1820696112513542
|
| 30 |
+
],
|
| 31 |
+
"q99": [
|
| 32 |
+
0.39412812441587447,
|
| 33 |
+
1.09063316822052,
|
| 34 |
+
0.1658253240585327
|
| 35 |
+
]
|
| 36 |
+
},
|
| 37 |
+
"eef_rotation": {
|
| 38 |
+
"max": [
|
| 39 |
+
-0.35259532928466797,
|
| 40 |
+
0.8778502345085144,
|
| 41 |
+
2.965937376022339
|
| 42 |
+
],
|
| 43 |
+
"min": [
|
| 44 |
+
-2.3446121215820312,
|
| 45 |
+
-0.47837626934051514,
|
| 46 |
+
1.515358328819275
|
| 47 |
+
],
|
| 48 |
+
"mean": [
|
| 49 |
+
-1.6130940914154053,
|
| 50 |
+
-0.06260889768600464,
|
| 51 |
+
2.2893757820129395
|
| 52 |
+
],
|
| 53 |
+
"std": [
|
| 54 |
+
0.3057422637939453,
|
| 55 |
+
0.13727501034736633,
|
| 56 |
+
0.18679431080818176
|
| 57 |
+
],
|
| 58 |
+
"q01": [
|
| 59 |
+
-2.151788368225098,
|
| 60 |
+
-0.348803762793541,
|
| 61 |
+
1.8247082424163819
|
| 62 |
+
],
|
| 63 |
+
"q99": [
|
| 64 |
+
-0.8719659841060661,
|
| 65 |
+
0.32717462211847265,
|
| 66 |
+
2.695855288505554
|
| 67 |
+
]
|
| 68 |
+
},
|
| 69 |
+
"gripper_position": {
|
| 70 |
+
"max": [
|
| 71 |
+
1.0660839080810547
|
| 72 |
+
],
|
| 73 |
+
"min": [
|
| 74 |
+
-0.862152636051178
|
| 75 |
+
],
|
| 76 |
+
"mean": [
|
| 77 |
+
0.08788274973630905
|
| 78 |
+
],
|
| 79 |
+
"std": [
|
| 80 |
+
0.2687181830406189
|
| 81 |
+
],
|
| 82 |
+
"q01": [
|
| 83 |
+
-0.6035337620973587
|
| 84 |
+
],
|
| 85 |
+
"q99": [
|
| 86 |
+
0.7679916465282439
|
| 87 |
+
]
|
| 88 |
+
}
|
| 89 |
+
},
|
| 90 |
+
"action": {
|
| 91 |
+
"eef_position_delta": {
|
| 92 |
+
"max": [
|
| 93 |
+
0.5925613045692444,
|
| 94 |
+
0.8723994493484497,
|
| 95 |
+
0.69416743516922
|
| 96 |
+
],
|
| 97 |
+
"min": [
|
| 98 |
+
-0.960589587688446,
|
| 99 |
+
-0.8221790194511414,
|
| 100 |
+
-0.8575763702392578
|
| 101 |
+
],
|
| 102 |
+
"mean": [
|
| 103 |
+
0.03402156010270119,
|
| 104 |
+
-0.007893561385571957,
|
| 105 |
+
0.0033362761605530977
|
| 106 |
+
],
|
| 107 |
+
"std": [
|
| 108 |
+
0.19854718446731567,
|
| 109 |
+
0.15367349982261658,
|
| 110 |
+
0.14636090397834778
|
| 111 |
+
],
|
| 112 |
+
"q01": [
|
| 113 |
+
-0.47471584677696227,
|
| 114 |
+
-0.4062542155385018,
|
| 115 |
+
-0.26967047452926635
|
| 116 |
+
],
|
| 117 |
+
"q99": [
|
| 118 |
+
0.4079366648197173,
|
| 119 |
+
0.34532205790281273,
|
| 120 |
+
0.4184985920786849
|
| 121 |
+
]
|
| 122 |
+
},
|
| 123 |
+
"eef_rotation_delta": {
|
| 124 |
+
"max": [
|
| 125 |
+
0.8240683674812317,
|
| 126 |
+
0.7224147915840149,
|
| 127 |
+
0.8612414598464966
|
| 128 |
+
],
|
| 129 |
+
"min": [
|
| 130 |
+
-0.8100513815879822,
|
| 131 |
+
-0.8826454877853394,
|
| 132 |
+
-0.9472323656082153
|
| 133 |
+
],
|
| 134 |
+
"mean": [
|
| 135 |
+
0.0014922608388587832,
|
| 136 |
+
-0.029764438048005104,
|
| 137 |
+
0.006220718380063772
|
| 138 |
+
],
|
| 139 |
+
"std": [
|
| 140 |
+
0.14293360710144043,
|
| 141 |
+
0.16134193539619446,
|
| 142 |
+
0.2356378585100174
|
| 143 |
+
],
|
| 144 |
+
"q01": [
|
| 145 |
+
-0.3907526931166649,
|
| 146 |
+
-0.4691345453262329,
|
| 147 |
+
-0.5505707603693009
|
| 148 |
+
],
|
| 149 |
+
"q99": [
|
| 150 |
+
0.37015672087669316,
|
| 151 |
+
0.39511197239160506,
|
| 152 |
+
0.573674650788307
|
| 153 |
+
]
|
| 154 |
+
},
|
| 155 |
+
"gripper_position": {
|
| 156 |
+
"max": [
|
| 157 |
+
1.0
|
| 158 |
+
],
|
| 159 |
+
"min": [
|
| 160 |
+
0.0
|
| 161 |
+
],
|
| 162 |
+
"mean": [
|
| 163 |
+
0.6008301973342896
|
| 164 |
+
],
|
| 165 |
+
"std": [
|
| 166 |
+
0.3934294879436493
|
| 167 |
+
],
|
| 168 |
+
"q01": [
|
| 169 |
+
0.07158590108156204
|
| 170 |
+
],
|
| 171 |
+
"q99": [
|
| 172 |
+
1.0
|
| 173 |
+
]
|
| 174 |
+
}
|
| 175 |
+
}
|
| 176 |
+
},
|
| 177 |
+
"modalities": {
|
| 178 |
+
"video": {
|
| 179 |
+
"exterior_image_1": {
|
| 180 |
+
"resolution": [
|
| 181 |
+
320,
|
| 182 |
+
180
|
| 183 |
+
],
|
| 184 |
+
"channels": 3,
|
| 185 |
+
"fps": 15.0
|
| 186 |
+
},
|
| 187 |
+
"exterior_image_2": {
|
| 188 |
+
"resolution": [
|
| 189 |
+
320,
|
| 190 |
+
180
|
| 191 |
+
],
|
| 192 |
+
"channels": 3,
|
| 193 |
+
"fps": 15.0
|
| 194 |
+
},
|
| 195 |
+
"wrist_image": {
|
| 196 |
+
"resolution": [
|
| 197 |
+
320,
|
| 198 |
+
180
|
| 199 |
+
],
|
| 200 |
+
"channels": 3,
|
| 201 |
+
"fps": 15.0
|
| 202 |
+
}
|
| 203 |
+
},
|
| 204 |
+
"state": {
|
| 205 |
+
"eef_position": {
|
| 206 |
+
"absolute": true,
|
| 207 |
+
"rotation_type": null,
|
| 208 |
+
"shape": [
|
| 209 |
+
3
|
| 210 |
+
],
|
| 211 |
+
"continuous": true
|
| 212 |
+
},
|
| 213 |
+
"eef_rotation": {
|
| 214 |
+
"absolute": true,
|
| 215 |
+
"rotation_type": "euler_angles_rpy",
|
| 216 |
+
"shape": [
|
| 217 |
+
3
|
| 218 |
+
],
|
| 219 |
+
"continuous": true
|
| 220 |
+
},
|
| 221 |
+
"gripper_position": {
|
| 222 |
+
"absolute": true,
|
| 223 |
+
"rotation_type": null,
|
| 224 |
+
"shape": [
|
| 225 |
+
1
|
| 226 |
+
],
|
| 227 |
+
"continuous": true
|
| 228 |
+
}
|
| 229 |
+
},
|
| 230 |
+
"action": {
|
| 231 |
+
"eef_position_delta": {
|
| 232 |
+
"absolute": false,
|
| 233 |
+
"rotation_type": null,
|
| 234 |
+
"shape": [
|
| 235 |
+
3
|
| 236 |
+
],
|
| 237 |
+
"continuous": true
|
| 238 |
+
},
|
| 239 |
+
"eef_rotation_delta": {
|
| 240 |
+
"absolute": false,
|
| 241 |
+
"rotation_type": "axis_angle",
|
| 242 |
+
"shape": [
|
| 243 |
+
3
|
| 244 |
+
],
|
| 245 |
+
"continuous": true
|
| 246 |
+
},
|
| 247 |
+
"gripper_position": {
|
| 248 |
+
"absolute": false,
|
| 249 |
+
"rotation_type": null,
|
| 250 |
+
"shape": [
|
| 251 |
+
1
|
| 252 |
+
],
|
| 253 |
+
"continuous": true
|
| 254 |
+
}
|
| 255 |
+
}
|
| 256 |
+
},
|
| 257 |
+
"embodiment_tag": "oxe_droid"
|
| 258 |
+
}
|
| 259 |
+
}
|
model-00001-of-00002.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:da5487b7f3546322ddae34bb07dccab41feae6c722565237dc73a0f96861062b
|
| 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:c1f2def2be205cda90f1f730f18164dec31b074b0c482d18949dc4a78c32cce4
|
| 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 +1,743 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": null,
|
| 3 |
+
"best_metric": null,
|
| 4 |
+
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 1.9267822736030829,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 1000,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.019267822736030827,
|
| 14 |
+
"grad_norm": 1.7207869291305542,
|
| 15 |
+
"learning_rate": 1.8e-05,
|
| 16 |
+
"loss": 0.2916,
|
| 17 |
+
"step": 10
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.038535645472061654,
|
| 21 |
+
"grad_norm": 0.542586088180542,
|
| 22 |
+
"learning_rate": 3.8e-05,
|
| 23 |
+
"loss": 0.1243,
|
| 24 |
+
"step": 20
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.057803468208092484,
|
| 28 |
+
"grad_norm": 0.7696737051010132,
|
| 29 |
+
"learning_rate": 5.8e-05,
|
| 30 |
+
"loss": 0.1138,
|
| 31 |
+
"step": 30
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.07707129094412331,
|
| 35 |
+
"grad_norm": 0.524152398109436,
|
| 36 |
+
"learning_rate": 7.800000000000001e-05,
|
| 37 |
+
"loss": 0.1041,
|
| 38 |
+
"step": 40
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.09633911368015415,
|
| 42 |
+
"grad_norm": 0.4883207380771637,
|
| 43 |
+
"learning_rate": 9.8e-05,
|
| 44 |
+
"loss": 0.0938,
|
| 45 |
+
"step": 50
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.11560693641618497,
|
| 49 |
+
"grad_norm": 0.6807201504707336,
|
| 50 |
+
"learning_rate": 9.997785653888835e-05,
|
| 51 |
+
"loss": 0.1036,
|
| 52 |
+
"step": 60
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.1348747591522158,
|
| 56 |
+
"grad_norm": 0.5879040360450745,
|
| 57 |
+
"learning_rate": 9.990133642141359e-05,
|
| 58 |
+
"loss": 0.0913,
|
| 59 |
+
"step": 70
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 0.15414258188824662,
|
| 63 |
+
"grad_norm": 0.2884378135204315,
|
| 64 |
+
"learning_rate": 9.977024992520602e-05,
|
| 65 |
+
"loss": 0.0937,
|
| 66 |
+
"step": 80
|
| 67 |
+
},
|
| 68 |
+
{
|
| 69 |
+
"epoch": 0.17341040462427745,
|
| 70 |
+
"grad_norm": 0.31237179040908813,
|
| 71 |
+
"learning_rate": 9.95847403914247e-05,
|
| 72 |
+
"loss": 0.0818,
|
| 73 |
+
"step": 90
|
| 74 |
+
},
|
| 75 |
+
{
|
| 76 |
+
"epoch": 0.1926782273603083,
|
| 77 |
+
"grad_norm": 0.5337116718292236,
|
| 78 |
+
"learning_rate": 9.934501067202117e-05,
|
| 79 |
+
"loss": 0.0871,
|
| 80 |
+
"step": 100
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.2119460500963391,
|
| 84 |
+
"grad_norm": 0.35781776905059814,
|
| 85 |
+
"learning_rate": 9.905132290792394e-05,
|
| 86 |
+
"loss": 0.0813,
|
| 87 |
+
"step": 110
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"epoch": 0.23121387283236994,
|
| 91 |
+
"grad_norm": 0.41506218910217285,
|
| 92 |
+
"learning_rate": 9.870399824239117e-05,
|
| 93 |
+
"loss": 0.0807,
|
| 94 |
+
"step": 120
|
| 95 |
+
},
|
| 96 |
+
{
|
| 97 |
+
"epoch": 0.2504816955684008,
|
| 98 |
+
"grad_norm": 0.2967628836631775,
|
| 99 |
+
"learning_rate": 9.830341646984521e-05,
|
| 100 |
+
"loss": 0.0824,
|
| 101 |
+
"step": 130
|
| 102 |
+
},
|
| 103 |
+
{
|
| 104 |
+
"epoch": 0.2697495183044316,
|
| 105 |
+
"grad_norm": 0.31782740354537964,
|
| 106 |
+
"learning_rate": 9.785001562057309e-05,
|
| 107 |
+
"loss": 0.0756,
|
| 108 |
+
"step": 140
|
| 109 |
+
},
|
| 110 |
+
{
|
| 111 |
+
"epoch": 0.28901734104046245,
|
| 112 |
+
"grad_norm": 0.34691065549850464,
|
| 113 |
+
"learning_rate": 9.734429148174675e-05,
|
| 114 |
+
"loss": 0.084,
|
| 115 |
+
"step": 150
|
| 116 |
+
},
|
| 117 |
+
{
|
| 118 |
+
"epoch": 0.30828516377649323,
|
| 119 |
+
"grad_norm": 0.48853012919425964,
|
| 120 |
+
"learning_rate": 9.6786797055287e-05,
|
| 121 |
+
"loss": 0.0832,
|
| 122 |
+
"step": 160
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.32755298651252407,
|
| 126 |
+
"grad_norm": 0.39761826395988464,
|
| 127 |
+
"learning_rate": 9.617814195316411e-05,
|
| 128 |
+
"loss": 0.0773,
|
| 129 |
+
"step": 170
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.3468208092485549,
|
| 133 |
+
"grad_norm": 0.41890355944633484,
|
| 134 |
+
"learning_rate": 9.551899173079607e-05,
|
| 135 |
+
"loss": 0.072,
|
| 136 |
+
"step": 180
|
| 137 |
+
},
|
| 138 |
+
{
|
| 139 |
+
"epoch": 0.36608863198458574,
|
| 140 |
+
"grad_norm": 0.5497244596481323,
|
| 141 |
+
"learning_rate": 9.481006715927351e-05,
|
| 142 |
+
"loss": 0.0733,
|
| 143 |
+
"step": 190
|
| 144 |
+
},
|
| 145 |
+
{
|
| 146 |
+
"epoch": 0.3853564547206166,
|
| 147 |
+
"grad_norm": 0.28932520747184753,
|
| 148 |
+
"learning_rate": 9.405214343720707e-05,
|
| 149 |
+
"loss": 0.0814,
|
| 150 |
+
"step": 200
|
| 151 |
+
},
|
| 152 |
+
{
|
| 153 |
+
"epoch": 0.4046242774566474,
|
| 154 |
+
"grad_norm": 0.5443249344825745,
|
| 155 |
+
"learning_rate": 9.32460493430591e-05,
|
| 156 |
+
"loss": 0.0789,
|
| 157 |
+
"step": 210
|
| 158 |
+
},
|
| 159 |
+
{
|
| 160 |
+
"epoch": 0.4238921001926782,
|
| 161 |
+
"grad_norm": 0.5669806003570557,
|
| 162 |
+
"learning_rate": 9.239266632888659e-05,
|
| 163 |
+
"loss": 0.0717,
|
| 164 |
+
"step": 220
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.44315992292870904,
|
| 168 |
+
"grad_norm": 0.3382496237754822,
|
| 169 |
+
"learning_rate": 9.14929275564863e-05,
|
| 170 |
+
"loss": 0.0714,
|
| 171 |
+
"step": 230
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 0.4624277456647399,
|
| 175 |
+
"grad_norm": 0.47498270869255066,
|
| 176 |
+
"learning_rate": 9.0547816876996e-05,
|
| 177 |
+
"loss": 0.0691,
|
| 178 |
+
"step": 240
|
| 179 |
+
},
|
| 180 |
+
{
|
| 181 |
+
"epoch": 0.4816955684007707,
|
| 182 |
+
"grad_norm": 0.4482809007167816,
|
| 183 |
+
"learning_rate": 8.955836775506776e-05,
|
| 184 |
+
"loss": 0.0694,
|
| 185 |
+
"step": 250
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"epoch": 0.5009633911368016,
|
| 189 |
+
"grad_norm": 0.5501825213432312,
|
| 190 |
+
"learning_rate": 8.852566213878947e-05,
|
| 191 |
+
"loss": 0.0686,
|
| 192 |
+
"step": 260
|
| 193 |
+
},
|
| 194 |
+
{
|
| 195 |
+
"epoch": 0.5202312138728323,
|
| 196 |
+
"grad_norm": 0.3987138867378235,
|
| 197 |
+
"learning_rate": 8.745082927659047e-05,
|
| 198 |
+
"loss": 0.078,
|
| 199 |
+
"step": 270
|
| 200 |
+
},
|
| 201 |
+
{
|
| 202 |
+
"epoch": 0.5394990366088632,
|
| 203 |
+
"grad_norm": 0.28167715668678284,
|
| 204 |
+
"learning_rate": 8.633504448242505e-05,
|
| 205 |
+
"loss": 0.0758,
|
| 206 |
+
"step": 280
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.558766859344894,
|
| 210 |
+
"grad_norm": 0.24536319077014923,
|
| 211 |
+
"learning_rate": 8.517952785058385e-05,
|
| 212 |
+
"loss": 0.0747,
|
| 213 |
+
"step": 290
|
| 214 |
+
},
|
| 215 |
+
{
|
| 216 |
+
"epoch": 0.5780346820809249,
|
| 217 |
+
"grad_norm": 0.2790259122848511,
|
| 218 |
+
"learning_rate": 8.398554292153866e-05,
|
| 219 |
+
"loss": 0.0681,
|
| 220 |
+
"step": 300
|
| 221 |
+
},
|
| 222 |
+
{
|
| 223 |
+
"epoch": 0.5973025048169557,
|
| 224 |
+
"grad_norm": 0.4405696988105774,
|
| 225 |
+
"learning_rate": 8.275439530027948e-05,
|
| 226 |
+
"loss": 0.0713,
|
| 227 |
+
"step": 310
|
| 228 |
+
},
|
| 229 |
+
{
|
| 230 |
+
"epoch": 0.6165703275529865,
|
| 231 |
+
"grad_norm": 0.3930000066757202,
|
| 232 |
+
"learning_rate": 8.148743122865463e-05,
|
| 233 |
+
"loss": 0.0633,
|
| 234 |
+
"step": 320
|
| 235 |
+
},
|
| 236 |
+
{
|
| 237 |
+
"epoch": 0.6358381502890174,
|
| 238 |
+
"grad_norm": 0.39871180057525635,
|
| 239 |
+
"learning_rate": 8.018603611327504e-05,
|
| 240 |
+
"loss": 0.0735,
|
| 241 |
+
"step": 330
|
| 242 |
+
},
|
| 243 |
+
{
|
| 244 |
+
"epoch": 0.6551059730250481,
|
| 245 |
+
"grad_norm": 0.43595343828201294,
|
| 246 |
+
"learning_rate": 7.88516330105925e-05,
|
| 247 |
+
"loss": 0.0651,
|
| 248 |
+
"step": 340
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 0.674373795761079,
|
| 252 |
+
"grad_norm": 0.3049931824207306,
|
| 253 |
+
"learning_rate": 7.748568107080832e-05,
|
| 254 |
+
"loss": 0.0641,
|
| 255 |
+
"step": 350
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 0.6936416184971098,
|
| 259 |
+
"grad_norm": 0.2661554515361786,
|
| 260 |
+
"learning_rate": 7.608967394231387e-05,
|
| 261 |
+
"loss": 0.0615,
|
| 262 |
+
"step": 360
|
| 263 |
+
},
|
| 264 |
+
{
|
| 265 |
+
"epoch": 0.7129094412331407,
|
| 266 |
+
"grad_norm": 0.23182271420955658,
|
| 267 |
+
"learning_rate": 7.466513813840825e-05,
|
| 268 |
+
"loss": 0.0641,
|
| 269 |
+
"step": 370
|
| 270 |
+
},
|
| 271 |
+
{
|
| 272 |
+
"epoch": 0.7321772639691715,
|
| 273 |
+
"grad_norm": 0.36455002427101135,
|
| 274 |
+
"learning_rate": 7.32136313680782e-05,
|
| 275 |
+
"loss": 0.0694,
|
| 276 |
+
"step": 380
|
| 277 |
+
},
|
| 278 |
+
{
|
| 279 |
+
"epoch": 0.7514450867052023,
|
| 280 |
+
"grad_norm": 0.37729915976524353,
|
| 281 |
+
"learning_rate": 7.173674083266624e-05,
|
| 282 |
+
"loss": 0.0657,
|
| 283 |
+
"step": 390
|
| 284 |
+
},
|
| 285 |
+
{
|
| 286 |
+
"epoch": 0.7707129094412332,
|
| 287 |
+
"grad_norm": 0.3024880290031433,
|
| 288 |
+
"learning_rate": 7.023608149028937e-05,
|
| 289 |
+
"loss": 0.0595,
|
| 290 |
+
"step": 400
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 0.789980732177264,
|
| 294 |
+
"grad_norm": 0.31104758381843567,
|
| 295 |
+
"learning_rate": 6.871329428990602e-05,
|
| 296 |
+
"loss": 0.0681,
|
| 297 |
+
"step": 410
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 0.8092485549132948,
|
| 301 |
+
"grad_norm": 0.350875586271286,
|
| 302 |
+
"learning_rate": 6.71700443769625e-05,
|
| 303 |
+
"loss": 0.061,
|
| 304 |
+
"step": 420
|
| 305 |
+
},
|
| 306 |
+
{
|
| 307 |
+
"epoch": 0.8285163776493256,
|
| 308 |
+
"grad_norm": 0.2729007303714752,
|
| 309 |
+
"learning_rate": 6.56080192725808e-05,
|
| 310 |
+
"loss": 0.0666,
|
| 311 |
+
"step": 430
|
| 312 |
+
},
|
| 313 |
+
{
|
| 314 |
+
"epoch": 0.8477842003853564,
|
| 315 |
+
"grad_norm": 0.29519274830818176,
|
| 316 |
+
"learning_rate": 6.402892702827916e-05,
|
| 317 |
+
"loss": 0.0617,
|
| 318 |
+
"step": 440
|
| 319 |
+
},
|
| 320 |
+
{
|
| 321 |
+
"epoch": 0.8670520231213873,
|
| 322 |
+
"grad_norm": 0.41105949878692627,
|
| 323 |
+
"learning_rate": 6.243449435824276e-05,
|
| 324 |
+
"loss": 0.0635,
|
| 325 |
+
"step": 450
|
| 326 |
+
},
|
| 327 |
+
{
|
| 328 |
+
"epoch": 0.8863198458574181,
|
| 329 |
+
"grad_norm": 0.39932650327682495,
|
| 330 |
+
"learning_rate": 6.0826464751186994e-05,
|
| 331 |
+
"loss": 0.0668,
|
| 332 |
+
"step": 460
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 0.905587668593449,
|
| 336 |
+
"grad_norm": 0.36617910861968994,
|
| 337 |
+
"learning_rate": 5.9206596563878357e-05,
|
| 338 |
+
"loss": 0.0642,
|
| 339 |
+
"step": 470
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 0.9248554913294798,
|
| 343 |
+
"grad_norm": 0.3059275150299072,
|
| 344 |
+
"learning_rate": 5.757666109839702e-05,
|
| 345 |
+
"loss": 0.0639,
|
| 346 |
+
"step": 480
|
| 347 |
+
},
|
| 348 |
+
{
|
| 349 |
+
"epoch": 0.9441233140655106,
|
| 350 |
+
"grad_norm": 0.3001484274864197,
|
| 351 |
+
"learning_rate": 5.5938440665244006e-05,
|
| 352 |
+
"loss": 0.0581,
|
| 353 |
+
"step": 490
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"epoch": 0.9633911368015414,
|
| 357 |
+
"grad_norm": 0.3571004569530487,
|
| 358 |
+
"learning_rate": 5.4293726634410855e-05,
|
| 359 |
+
"loss": 0.0558,
|
| 360 |
+
"step": 500
|
| 361 |
+
},
|
| 362 |
+
{
|
| 363 |
+
"epoch": 0.9826589595375722,
|
| 364 |
+
"grad_norm": 0.3813907206058502,
|
| 365 |
+
"learning_rate": 5.264431747654284e-05,
|
| 366 |
+
"loss": 0.0618,
|
| 367 |
+
"step": 510
|
| 368 |
+
},
|
| 369 |
+
{
|
| 370 |
+
"epoch": 1.001926782273603,
|
| 371 |
+
"grad_norm": 0.29777514934539795,
|
| 372 |
+
"learning_rate": 5.0992016796337686e-05,
|
| 373 |
+
"loss": 0.0578,
|
| 374 |
+
"step": 520
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 1.0211946050096339,
|
| 378 |
+
"grad_norm": 0.28990840911865234,
|
| 379 |
+
"learning_rate": 4.93386313603304e-05,
|
| 380 |
+
"loss": 0.0613,
|
| 381 |
+
"step": 530
|
| 382 |
+
},
|
| 383 |
+
{
|
| 384 |
+
"epoch": 1.0404624277456647,
|
| 385 |
+
"grad_norm": 0.43356871604919434,
|
| 386 |
+
"learning_rate": 4.7685969121220456e-05,
|
| 387 |
+
"loss": 0.057,
|
| 388 |
+
"step": 540
|
| 389 |
+
},
|
| 390 |
+
{
|
| 391 |
+
"epoch": 1.0597302504816957,
|
| 392 |
+
"grad_norm": 0.2468729466199875,
|
| 393 |
+
"learning_rate": 4.60358372409022e-05,
|
| 394 |
+
"loss": 0.053,
|
| 395 |
+
"step": 550
|
| 396 |
+
},
|
| 397 |
+
{
|
| 398 |
+
"epoch": 1.0789980732177264,
|
| 399 |
+
"grad_norm": 0.33707353472709656,
|
| 400 |
+
"learning_rate": 4.439004011435979e-05,
|
| 401 |
+
"loss": 0.0523,
|
| 402 |
+
"step": 560
|
| 403 |
+
},
|
| 404 |
+
{
|
| 405 |
+
"epoch": 1.0982658959537572,
|
| 406 |
+
"grad_norm": 0.3425613343715668,
|
| 407 |
+
"learning_rate": 4.275037739658771e-05,
|
| 408 |
+
"loss": 0.0533,
|
| 409 |
+
"step": 570
|
| 410 |
+
},
|
| 411 |
+
{
|
| 412 |
+
"epoch": 1.117533718689788,
|
| 413 |
+
"grad_norm": 0.21212391555309296,
|
| 414 |
+
"learning_rate": 4.111864203469457e-05,
|
| 415 |
+
"loss": 0.0506,
|
| 416 |
+
"step": 580
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 1.1368015414258188,
|
| 420 |
+
"grad_norm": 0.2778816223144531,
|
| 421 |
+
"learning_rate": 3.949661830734172e-05,
|
| 422 |
+
"loss": 0.0575,
|
| 423 |
+
"step": 590
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"epoch": 1.1560693641618498,
|
| 427 |
+
"grad_norm": 0.2917283773422241,
|
| 428 |
+
"learning_rate": 3.788607987366069e-05,
|
| 429 |
+
"loss": 0.0571,
|
| 430 |
+
"step": 600
|
| 431 |
+
},
|
| 432 |
+
{
|
| 433 |
+
"epoch": 1.1753371868978806,
|
| 434 |
+
"grad_norm": 0.27486225962638855,
|
| 435 |
+
"learning_rate": 3.628878783378302e-05,
|
| 436 |
+
"loss": 0.056,
|
| 437 |
+
"step": 610
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 1.1946050096339114,
|
| 441 |
+
"grad_norm": 0.40446701645851135,
|
| 442 |
+
"learning_rate": 3.470648880310313e-05,
|
| 443 |
+
"loss": 0.0529,
|
| 444 |
+
"step": 620
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 1.2138728323699421,
|
| 448 |
+
"grad_norm": 0.25892701745033264,
|
| 449 |
+
"learning_rate": 3.3140913002379995e-05,
|
| 450 |
+
"loss": 0.0504,
|
| 451 |
+
"step": 630
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 1.2331406551059731,
|
| 455 |
+
"grad_norm": 0.26697343587875366,
|
| 456 |
+
"learning_rate": 3.1593772365766105e-05,
|
| 457 |
+
"loss": 0.0509,
|
| 458 |
+
"step": 640
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 1.252408477842004,
|
| 462 |
+
"grad_norm": 0.2691023051738739,
|
| 463 |
+
"learning_rate": 3.006675866883275e-05,
|
| 464 |
+
"loss": 0.0576,
|
| 465 |
+
"step": 650
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 1.2716763005780347,
|
| 469 |
+
"grad_norm": 0.2726045250892639,
|
| 470 |
+
"learning_rate": 2.8561541678638142e-05,
|
| 471 |
+
"loss": 0.051,
|
| 472 |
+
"step": 660
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 1.2909441233140655,
|
| 476 |
+
"grad_norm": 0.3852327764034271,
|
| 477 |
+
"learning_rate": 2.707976732786166e-05,
|
| 478 |
+
"loss": 0.0535,
|
| 479 |
+
"step": 670
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 1.3102119460500963,
|
| 483 |
+
"grad_norm": 0.2920765280723572,
|
| 484 |
+
"learning_rate": 2.562305591500069e-05,
|
| 485 |
+
"loss": 0.0492,
|
| 486 |
+
"step": 680
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 1.3294797687861273,
|
| 490 |
+
"grad_norm": 0.1913856863975525,
|
| 491 |
+
"learning_rate": 2.419300033259798e-05,
|
| 492 |
+
"loss": 0.0507,
|
| 493 |
+
"step": 690
|
| 494 |
+
},
|
| 495 |
+
{
|
| 496 |
+
"epoch": 1.348747591522158,
|
| 497 |
+
"grad_norm": 0.29561716318130493,
|
| 498 |
+
"learning_rate": 2.279116432543705e-05,
|
| 499 |
+
"loss": 0.0499,
|
| 500 |
+
"step": 700
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 1.3680154142581888,
|
| 504 |
+
"grad_norm": 0.4290018379688263,
|
| 505 |
+
"learning_rate": 2.1419080780610123e-05,
|
| 506 |
+
"loss": 0.0571,
|
| 507 |
+
"step": 710
|
| 508 |
+
},
|
| 509 |
+
{
|
| 510 |
+
"epoch": 1.3872832369942196,
|
| 511 |
+
"grad_norm": 0.2543991506099701,
|
| 512 |
+
"learning_rate": 2.0078250051328784e-05,
|
| 513 |
+
"loss": 0.0524,
|
| 514 |
+
"step": 720
|
| 515 |
+
},
|
| 516 |
+
{
|
| 517 |
+
"epoch": 1.4065510597302504,
|
| 518 |
+
"grad_norm": 0.2487826943397522,
|
| 519 |
+
"learning_rate": 1.877013831630961e-05,
|
| 520 |
+
"loss": 0.0539,
|
| 521 |
+
"step": 730
|
| 522 |
+
},
|
| 523 |
+
{
|
| 524 |
+
"epoch": 1.4258188824662814,
|
| 525 |
+
"grad_norm": 0.3550814390182495,
|
| 526 |
+
"learning_rate": 1.749617597652934e-05,
|
| 527 |
+
"loss": 0.0436,
|
| 528 |
+
"step": 740
|
| 529 |
+
},
|
| 530 |
+
{
|
| 531 |
+
"epoch": 1.4450867052023122,
|
| 532 |
+
"grad_norm": 0.2873481810092926,
|
| 533 |
+
"learning_rate": 1.62577560911024e-05,
|
| 534 |
+
"loss": 0.0463,
|
| 535 |
+
"step": 750
|
| 536 |
+
},
|
| 537 |
+
{
|
| 538 |
+
"epoch": 1.464354527938343,
|
| 539 |
+
"grad_norm": 0.22322338819503784,
|
| 540 |
+
"learning_rate": 1.5056232853991209e-05,
|
| 541 |
+
"loss": 0.0427,
|
| 542 |
+
"step": 760
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 1.4836223506743738,
|
| 546 |
+
"grad_norm": 0.23233197629451752,
|
| 547 |
+
"learning_rate": 1.389292011321498e-05,
|
| 548 |
+
"loss": 0.0427,
|
| 549 |
+
"step": 770
|
| 550 |
+
},
|
| 551 |
+
{
|
| 552 |
+
"epoch": 1.5028901734104045,
|
| 553 |
+
"grad_norm": 0.25211092829704285,
|
| 554 |
+
"learning_rate": 1.2769089934176126e-05,
|
| 555 |
+
"loss": 0.0456,
|
| 556 |
+
"step": 780
|
| 557 |
+
},
|
| 558 |
+
{
|
| 559 |
+
"epoch": 1.5221579961464355,
|
| 560 |
+
"grad_norm": 0.19579973816871643,
|
| 561 |
+
"learning_rate": 1.1685971208675539e-05,
|
| 562 |
+
"loss": 0.0447,
|
| 563 |
+
"step": 790
|
| 564 |
+
},
|
| 565 |
+
{
|
| 566 |
+
"epoch": 1.5414258188824663,
|
| 567 |
+
"grad_norm": 0.2609504461288452,
|
| 568 |
+
"learning_rate": 1.0644748311137376e-05,
|
| 569 |
+
"loss": 0.0467,
|
| 570 |
+
"step": 800
|
| 571 |
+
},
|
| 572 |
+
{
|
| 573 |
+
"epoch": 1.560693641618497,
|
| 574 |
+
"grad_norm": 0.23123040795326233,
|
| 575 |
+
"learning_rate": 9.646559803512994e-06,
|
| 576 |
+
"loss": 0.0505,
|
| 577 |
+
"step": 810
|
| 578 |
+
},
|
| 579 |
+
{
|
| 580 |
+
"epoch": 1.579961464354528,
|
| 581 |
+
"grad_norm": 0.1902703195810318,
|
| 582 |
+
"learning_rate": 8.692497190280224e-06,
|
| 583 |
+
"loss": 0.0463,
|
| 584 |
+
"step": 820
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 1.5992292870905587,
|
| 588 |
+
"grad_norm": 0.30633774399757385,
|
| 589 |
+
"learning_rate": 7.783603724899257e-06,
|
| 590 |
+
"loss": 0.0465,
|
| 591 |
+
"step": 830
|
| 592 |
+
},
|
| 593 |
+
{
|
| 594 |
+
"epoch": 1.6184971098265897,
|
| 595 |
+
"grad_norm": 0.24036064743995667,
|
| 596 |
+
"learning_rate": 6.92087326903022e-06,
|
| 597 |
+
"loss": 0.0435,
|
| 598 |
+
"step": 840
|
| 599 |
+
},
|
| 600 |
+
{
|
| 601 |
+
"epoch": 1.6377649325626205,
|
| 602 |
+
"grad_norm": 0.20757320523262024,
|
| 603 |
+
"learning_rate": 6.1052492057601275e-06,
|
| 604 |
+
"loss": 0.0505,
|
| 605 |
+
"step": 850
|
| 606 |
+
},
|
| 607 |
+
{
|
| 608 |
+
"epoch": 1.6570327552986512,
|
| 609 |
+
"grad_norm": 0.3485427796840668,
|
| 610 |
+
"learning_rate": 5.337623408027293e-06,
|
| 611 |
+
"loss": 0.0441,
|
| 612 |
+
"step": 860
|
| 613 |
+
},
|
| 614 |
+
{
|
| 615 |
+
"epoch": 1.6763005780346822,
|
| 616 |
+
"grad_norm": 0.235531747341156,
|
| 617 |
+
"learning_rate": 4.618835263371396e-06,
|
| 618 |
+
"loss": 0.0464,
|
| 619 |
+
"step": 870
|
| 620 |
+
},
|
| 621 |
+
{
|
| 622 |
+
"epoch": 1.6955684007707128,
|
| 623 |
+
"grad_norm": 0.19758078455924988,
|
| 624 |
+
"learning_rate": 3.949670756075447e-06,
|
| 625 |
+
"loss": 0.0413,
|
| 626 |
+
"step": 880
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 1.7148362235067438,
|
| 630 |
+
"grad_norm": 0.21413107216358185,
|
| 631 |
+
"learning_rate": 3.3308616077036115e-06,
|
| 632 |
+
"loss": 0.0412,
|
| 633 |
+
"step": 890
|
| 634 |
+
},
|
| 635 |
+
{
|
| 636 |
+
"epoch": 1.7341040462427746,
|
| 637 |
+
"grad_norm": 0.26742419600486755,
|
| 638 |
+
"learning_rate": 2.7630844769743757e-06,
|
| 639 |
+
"loss": 0.0454,
|
| 640 |
+
"step": 900
|
| 641 |
+
},
|
| 642 |
+
{
|
| 643 |
+
"epoch": 1.7533718689788054,
|
| 644 |
+
"grad_norm": 0.22228221595287323,
|
| 645 |
+
"learning_rate": 2.2469602198441573e-06,
|
| 646 |
+
"loss": 0.0392,
|
| 647 |
+
"step": 910
|
| 648 |
+
},
|
| 649 |
+
{
|
| 650 |
+
"epoch": 1.7726396917148364,
|
| 651 |
+
"grad_norm": 0.2834679186344147,
|
| 652 |
+
"learning_rate": 1.7830532106104747e-06,
|
| 653 |
+
"loss": 0.0436,
|
| 654 |
+
"step": 920
|
| 655 |
+
},
|
| 656 |
+
{
|
| 657 |
+
"epoch": 1.791907514450867,
|
| 658 |
+
"grad_norm": 0.260470449924469,
|
| 659 |
+
"learning_rate": 1.3718707247769135e-06,
|
| 660 |
+
"loss": 0.0424,
|
| 661 |
+
"step": 930
|
| 662 |
+
},
|
| 663 |
+
{
|
| 664 |
+
"epoch": 1.811175337186898,
|
| 665 |
+
"grad_norm": 0.2204512655735016,
|
| 666 |
+
"learning_rate": 1.0138623843548078e-06,
|
| 667 |
+
"loss": 0.0393,
|
| 668 |
+
"step": 940
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 1.8304431599229287,
|
| 672 |
+
"grad_norm": 0.23786160349845886,
|
| 673 |
+
"learning_rate": 7.094196662081831e-07,
|
| 674 |
+
"loss": 0.0472,
|
| 675 |
+
"step": 950
|
| 676 |
+
},
|
| 677 |
+
{
|
| 678 |
+
"epoch": 1.8497109826589595,
|
| 679 |
+
"grad_norm": 0.18833257257938385,
|
| 680 |
+
"learning_rate": 4.5887547397955864e-07,
|
| 681 |
+
"loss": 0.0416,
|
| 682 |
+
"step": 960
|
| 683 |
+
},
|
| 684 |
+
{
|
| 685 |
+
"epoch": 1.8689788053949905,
|
| 686 |
+
"grad_norm": 0.20266398787498474,
|
| 687 |
+
"learning_rate": 2.625037740646763e-07,
|
| 688 |
+
"loss": 0.0426,
|
| 689 |
+
"step": 970
|
| 690 |
+
},
|
| 691 |
+
{
|
| 692 |
+
"epoch": 1.888246628131021,
|
| 693 |
+
"grad_norm": 0.2135821133852005,
|
| 694 |
+
"learning_rate": 1.2051929603428825e-07,
|
| 695 |
+
"loss": 0.0445,
|
| 696 |
+
"step": 980
|
| 697 |
+
},
|
| 698 |
+
{
|
| 699 |
+
"epoch": 1.907514450867052,
|
| 700 |
+
"grad_norm": 0.2030361145734787,
|
| 701 |
+
"learning_rate": 3.3077297830541584e-08,
|
| 702 |
+
"loss": 0.045,
|
| 703 |
+
"step": 990
|
| 704 |
+
},
|
| 705 |
+
{
|
| 706 |
+
"epoch": 1.9267822736030829,
|
| 707 |
+
"grad_norm": 0.31092095375061035,
|
| 708 |
+
"learning_rate": 2.7339599464326627e-10,
|
| 709 |
+
"loss": 0.0434,
|
| 710 |
+
"step": 1000
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 1.9267822736030829,
|
| 714 |
+
"step": 1000,
|
| 715 |
+
"total_flos": 0.0,
|
| 716 |
+
"train_loss": 0.06433489021658897,
|
| 717 |
+
"train_runtime": 935.0208,
|
| 718 |
+
"train_samples_per_second": 34.224,
|
| 719 |
+
"train_steps_per_second": 1.069
|
| 720 |
+
}
|
| 721 |
+
],
|
| 722 |
+
"logging_steps": 10,
|
| 723 |
+
"max_steps": 1000,
|
| 724 |
+
"num_input_tokens_seen": 0,
|
| 725 |
+
"num_train_epochs": 2,
|
| 726 |
+
"save_steps": 1000,
|
| 727 |
+
"stateful_callbacks": {
|
| 728 |
+
"TrainerControl": {
|
| 729 |
+
"args": {
|
| 730 |
+
"should_epoch_stop": false,
|
| 731 |
+
"should_evaluate": false,
|
| 732 |
+
"should_log": false,
|
| 733 |
+
"should_save": true,
|
| 734 |
+
"should_training_stop": true
|
| 735 |
+
},
|
| 736 |
+
"attributes": {}
|
| 737 |
+
}
|
| 738 |
+
},
|
| 739 |
+
"total_flos": 0.0,
|
| 740 |
+
"train_batch_size": 32,
|
| 741 |
+
"trial_name": null,
|
| 742 |
+
"trial_params": null
|
| 743 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f7333e1af2bdf6aa8ec2a60059ce66a628f29eb126dae4e4e6a1b469b2b0715e
|
| 3 |
+
size 5304
|