Run 8. Outer Step 33. Inner Step 4.
Browse files- config.json +215 -224
- generation_config.json +7 -0
- gradients.pt +3 -0
- inner_optimizer.pt +2 -2
- model.safetensors +2 -2
config.json
CHANGED
|
@@ -1,302 +1,293 @@
|
|
| 1 |
{
|
| 2 |
-
"_name_or_path": "distributed/
|
| 3 |
-
"activation_function": "gelu_new",
|
| 4 |
"all_reduce_scores": {
|
| 5 |
-
"0": "
|
| 6 |
"1": "SUCCESS",
|
| 7 |
-
"10": "
|
| 8 |
-
"100": "
|
| 9 |
-
"101": "
|
| 10 |
"102": "NON_PARTICIPATING",
|
| 11 |
-
"103": "
|
| 12 |
-
"104": "
|
| 13 |
"105": "SUCCESS",
|
| 14 |
-
"106": "
|
| 15 |
-
"107": "
|
| 16 |
-
"108": "
|
| 17 |
-
"109": "
|
| 18 |
-
"11": "
|
| 19 |
"110": "NON_PARTICIPATING",
|
| 20 |
-
"111": "
|
| 21 |
-
"112": "
|
| 22 |
-
"113": "
|
| 23 |
-
"114": "
|
| 24 |
-
"115": "
|
| 25 |
-
"116": "
|
| 26 |
-
"117": "
|
| 27 |
-
"118": "
|
| 28 |
-
"119": "
|
| 29 |
-
"12": "
|
| 30 |
-
"120": "
|
| 31 |
"121": "SUCCESS",
|
| 32 |
-
"122": "
|
| 33 |
-
"123": "
|
| 34 |
-
"124": "
|
| 35 |
-
"125": "
|
| 36 |
"126": "NON_PARTICIPATING",
|
| 37 |
-
"127": "
|
| 38 |
-
"128": "
|
| 39 |
-
"129": "
|
| 40 |
-
"13": "
|
| 41 |
-
"130": "
|
| 42 |
"131": "NON_PARTICIPATING",
|
| 43 |
-
"132": "
|
| 44 |
-
"133": "
|
| 45 |
-
"134": "
|
| 46 |
-
"135": "
|
| 47 |
"136": "NON_PARTICIPATING",
|
| 48 |
"137": "NON_PARTICIPATING",
|
| 49 |
-
"138": "
|
| 50 |
-
"139": "
|
| 51 |
"14": "SUCCESS",
|
| 52 |
"140": "SUCCESS",
|
| 53 |
-
"141": "
|
| 54 |
"142": "NON_PARTICIPATING",
|
| 55 |
-
"143": "
|
| 56 |
-
"144": "
|
| 57 |
-
"145": "
|
| 58 |
-
"146": "
|
| 59 |
-
"147": "
|
| 60 |
"148": "SUCCESS",
|
| 61 |
-
"149": "
|
| 62 |
"15": "SUCCESS",
|
| 63 |
-
"150": "
|
| 64 |
-
"151": "
|
| 65 |
-
"152": "
|
| 66 |
-
"153": "
|
| 67 |
"154": "NON_PARTICIPATING",
|
| 68 |
-
"155": "
|
| 69 |
-
"156": "
|
| 70 |
"157": "NON_PARTICIPATING",
|
| 71 |
"158": "NON_PARTICIPATING",
|
| 72 |
"159": "NON_PARTICIPATING",
|
| 73 |
"16": "SUCCESS",
|
| 74 |
"160": "NON_PARTICIPATING",
|
| 75 |
-
"161": "
|
| 76 |
-
"162": "
|
| 77 |
-
"163": "
|
| 78 |
"164": "NON_PARTICIPATING",
|
| 79 |
"165": "NON_PARTICIPATING",
|
| 80 |
"166": "SUCCESS",
|
| 81 |
-
"167": "
|
| 82 |
"168": "NON_PARTICIPATING",
|
| 83 |
"169": "NON_PARTICIPATING",
|
| 84 |
-
"17": "
|
| 85 |
"170": "NON_PARTICIPATING",
|
| 86 |
"171": "NON_PARTICIPATING",
|
| 87 |
-
"172": "
|
| 88 |
-
"173": "
|
| 89 |
-
"174": "
|
| 90 |
-
"175": "
|
| 91 |
-
"176": "
|
| 92 |
-
"177": "
|
| 93 |
-
"178": "
|
| 94 |
-
"179": "
|
| 95 |
-
"18": "
|
| 96 |
"180": "NON_PARTICIPATING",
|
| 97 |
"181": "SUCCESS",
|
| 98 |
-
"182": "
|
| 99 |
-
"183": "
|
| 100 |
-
"184": "
|
| 101 |
-
"185": "
|
| 102 |
-
"186": "
|
| 103 |
-
"187": "
|
| 104 |
-
"188": "
|
| 105 |
-
"189": "
|
| 106 |
-
"19": "
|
| 107 |
-
"190": "
|
| 108 |
"191": "NON_PARTICIPATING",
|
| 109 |
-
"192": "
|
| 110 |
-
"193": "
|
| 111 |
-
"194": "
|
| 112 |
-
"195": "
|
| 113 |
-
"196": "
|
| 114 |
-
"197": "
|
| 115 |
"198": "NON_PARTICIPATING",
|
| 116 |
-
"199": "
|
| 117 |
-
"2": "
|
| 118 |
-
"20": "
|
| 119 |
-
"200": "
|
| 120 |
-
"201": "
|
| 121 |
"202": "NON_PARTICIPATING",
|
| 122 |
-
"203": "
|
| 123 |
-
"204": "
|
| 124 |
-
"205": "
|
| 125 |
"206": "NON_PARTICIPATING",
|
| 126 |
-
"207": "
|
| 127 |
-
"208": "
|
| 128 |
-
"209": "
|
| 129 |
"21": "SUCCESS",
|
| 130 |
-
"210": "
|
| 131 |
-
"211": "
|
| 132 |
"212": "NON_PARTICIPATING",
|
| 133 |
"213": "NON_PARTICIPATING",
|
| 134 |
-
"214": "
|
| 135 |
-
"215": "
|
| 136 |
-
"216": "
|
| 137 |
-
"217": "
|
| 138 |
"218": "NON_PARTICIPATING",
|
| 139 |
-
"219": "
|
| 140 |
-
"22": "
|
| 141 |
"220": "NON_PARTICIPATING",
|
| 142 |
"221": "NON_PARTICIPATING",
|
| 143 |
-
"222": "
|
| 144 |
-
"223": "
|
| 145 |
"224": "NON_PARTICIPATING",
|
| 146 |
-
"225": "
|
| 147 |
-
"226": "
|
| 148 |
-
"227": "
|
| 149 |
-
"228": "
|
| 150 |
"229": "NON_PARTICIPATING",
|
| 151 |
-
"23": "
|
| 152 |
-
"230": "
|
| 153 |
"231": "NON_PARTICIPATING",
|
| 154 |
-
"232": "
|
| 155 |
"233": "NON_PARTICIPATING",
|
| 156 |
-
"234": "
|
| 157 |
-
"235": "
|
| 158 |
"236": "NON_PARTICIPATING",
|
| 159 |
-
"237": "
|
| 160 |
-
"238": "
|
| 161 |
"239": "NON_PARTICIPATING",
|
| 162 |
"24": "SUCCESS",
|
| 163 |
"240": "NON_PARTICIPATING",
|
| 164 |
"241": "NON_PARTICIPATING",
|
| 165 |
"242": "NON_PARTICIPATING",
|
| 166 |
-
"243": "
|
| 167 |
"244": "NON_PARTICIPATING",
|
| 168 |
-
"245": "
|
| 169 |
"246": "SUCCESS",
|
| 170 |
"247": "NON_PARTICIPATING",
|
| 171 |
-
"248": "
|
| 172 |
-
"249": "
|
| 173 |
-
"25": "
|
| 174 |
"250": "NON_PARTICIPATING",
|
| 175 |
-
"251": "
|
| 176 |
-
"252": "
|
| 177 |
-
"253": "
|
| 178 |
-
"254": "
|
| 179 |
"255": "NON_PARTICIPATING",
|
| 180 |
-
"26": "
|
| 181 |
"27": "SUCCESS",
|
| 182 |
"28": "SUCCESS",
|
| 183 |
-
"29": "
|
| 184 |
"3": "SUCCESS",
|
| 185 |
-
"30": "
|
| 186 |
-
"31": "
|
| 187 |
-
"32": "
|
| 188 |
-
"33": "
|
| 189 |
-
"34": "
|
| 190 |
-
"35": "
|
| 191 |
-
"36": "
|
| 192 |
-
"37": "
|
| 193 |
-
"38": "
|
| 194 |
-
"39": "
|
| 195 |
-
"4": "
|
| 196 |
-
"40": "
|
| 197 |
-
"41": "
|
| 198 |
-
"42": "
|
| 199 |
-
"43": "
|
| 200 |
"44": "SUCCESS",
|
| 201 |
-
"45": "
|
| 202 |
"46": "SUCCESS",
|
| 203 |
"47": "SUCCESS",
|
| 204 |
-
"48": "
|
| 205 |
-
"49": "
|
| 206 |
-
"5": "
|
| 207 |
-
"50": "
|
| 208 |
-
"51": "
|
| 209 |
"52": "SUCCESS",
|
| 210 |
-
"53": "
|
| 211 |
-
"54": "
|
| 212 |
-
"55": "
|
| 213 |
-
"56": "
|
| 214 |
-
"57": "
|
| 215 |
-
"58": "
|
| 216 |
-
"59": "
|
| 217 |
"6": "SUCCESS",
|
| 218 |
-
"60": "
|
| 219 |
-
"61": "
|
| 220 |
-
"62": "
|
| 221 |
-
"63": "
|
| 222 |
"64": "SUCCESS",
|
| 223 |
-
"65": "
|
| 224 |
-
"66": "
|
| 225 |
-
"67": "
|
| 226 |
-
"68": "
|
| 227 |
-
"69": "
|
| 228 |
-
"7": "
|
| 229 |
-
"70": "
|
| 230 |
-
"71": "
|
| 231 |
-
"72": "
|
| 232 |
"73": "SUCCESS",
|
| 233 |
-
"74": "
|
| 234 |
-
"75": "
|
| 235 |
-
"76": "
|
| 236 |
"77": "SUCCESS",
|
| 237 |
-
"78": "
|
| 238 |
-
"79": "
|
| 239 |
-
"8": "
|
| 240 |
-
"80": "
|
| 241 |
-
"81": "
|
| 242 |
-
"82": "
|
| 243 |
-
"83": "
|
| 244 |
-
"84": "
|
| 245 |
-
"85": "
|
| 246 |
-
"86": "
|
| 247 |
-
"87": "
|
| 248 |
"88": "SUCCESS",
|
| 249 |
-
"89": "
|
| 250 |
-
"9": "
|
| 251 |
-
"90": "
|
| 252 |
-
"91": "
|
| 253 |
-
"92": "
|
| 254 |
-
"93": "
|
| 255 |
-
"94": "
|
| 256 |
-
"95": "
|
| 257 |
-
"96": "
|
| 258 |
-
"97": "
|
| 259 |
"98": "SUCCESS",
|
| 260 |
-
"99": "
|
| 261 |
},
|
| 262 |
"architectures": [
|
| 263 |
-
"
|
| 264 |
],
|
| 265 |
-
"
|
| 266 |
-
"
|
| 267 |
-
"AutoConfig": "distributed/optimized-gpt2-500m--configuration_gpt_optimized.GPTOptimConfig",
|
| 268 |
-
"AutoModelForCausalLM": "distributed/optimized-gpt2-500m--modeling_gpt_optimized.GPTOptim"
|
| 269 |
-
},
|
| 270 |
"block_list": [
|
| 271 |
-
|
| 272 |
-
|
| 273 |
],
|
| 274 |
-
"
|
| 275 |
-
"
|
| 276 |
-
"
|
| 277 |
-
"
|
| 278 |
"initializer_range": 0.02,
|
| 279 |
-
"inner_step":
|
| 280 |
-
"
|
| 281 |
-
"last_allreduce_block":
|
| 282 |
-
"
|
| 283 |
-
"
|
| 284 |
-
"
|
| 285 |
-
"
|
| 286 |
-
"
|
| 287 |
-
"
|
| 288 |
-
"
|
| 289 |
-
"
|
| 290 |
-
"
|
| 291 |
-
"
|
| 292 |
-
"
|
| 293 |
-
"summary_activation": null,
|
| 294 |
-
"summary_first_dropout": 0.1,
|
| 295 |
-
"summary_proj_to_labels": true,
|
| 296 |
-
"summary_type": "cls_index",
|
| 297 |
-
"summary_use_proj": true,
|
| 298 |
"torch_dtype": "float32",
|
| 299 |
"transformers_version": "4.39.3",
|
| 300 |
-
"use_cache":
|
| 301 |
-
"vocab_size":
|
| 302 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"_name_or_path": "distributed/llama-1b",
|
|
|
|
| 3 |
"all_reduce_scores": {
|
| 4 |
+
"0": "NOT_ALIVE",
|
| 5 |
"1": "SUCCESS",
|
| 6 |
+
"10": "NOT_ALIVE",
|
| 7 |
+
"100": "NOT_ALIVE",
|
| 8 |
+
"101": "SUCCESS",
|
| 9 |
"102": "NON_PARTICIPATING",
|
| 10 |
+
"103": "NOT_ALIVE",
|
| 11 |
+
"104": "NOT_ALIVE",
|
| 12 |
"105": "SUCCESS",
|
| 13 |
+
"106": "FAIL",
|
| 14 |
+
"107": "NOT_ALIVE",
|
| 15 |
+
"108": "NOT_ALIVE",
|
| 16 |
+
"109": "NOT_ALIVE",
|
| 17 |
+
"11": "NOT_ALIVE",
|
| 18 |
"110": "NON_PARTICIPATING",
|
| 19 |
+
"111": "SUCCESS",
|
| 20 |
+
"112": "SUCCESS",
|
| 21 |
+
"113": "SUCCESS",
|
| 22 |
+
"114": "NOT_ALIVE",
|
| 23 |
+
"115": "SUCCESS",
|
| 24 |
+
"116": "NOT_ALIVE",
|
| 25 |
+
"117": "SUCCESS",
|
| 26 |
+
"118": "SUCCESS",
|
| 27 |
+
"119": "NOT_ALIVE",
|
| 28 |
+
"12": "NOT_ALIVE",
|
| 29 |
+
"120": "NOT_ALIVE",
|
| 30 |
"121": "SUCCESS",
|
| 31 |
+
"122": "NOT_ALIVE",
|
| 32 |
+
"123": "SUCCESS",
|
| 33 |
+
"124": "NOT_ALIVE",
|
| 34 |
+
"125": "NOT_ALIVE",
|
| 35 |
"126": "NON_PARTICIPATING",
|
| 36 |
+
"127": "NOT_ALIVE",
|
| 37 |
+
"128": "NOT_ALIVE",
|
| 38 |
+
"129": "SUCCESS",
|
| 39 |
+
"13": "NOT_ALIVE",
|
| 40 |
+
"130": "SUCCESS",
|
| 41 |
"131": "NON_PARTICIPATING",
|
| 42 |
+
"132": "SUCCESS",
|
| 43 |
+
"133": "SUCCESS",
|
| 44 |
+
"134": "SUCCESS",
|
| 45 |
+
"135": "SUCCESS",
|
| 46 |
"136": "NON_PARTICIPATING",
|
| 47 |
"137": "NON_PARTICIPATING",
|
| 48 |
+
"138": "SUCCESS",
|
| 49 |
+
"139": "NOT_ALIVE",
|
| 50 |
"14": "SUCCESS",
|
| 51 |
"140": "SUCCESS",
|
| 52 |
+
"141": "NOT_ALIVE",
|
| 53 |
"142": "NON_PARTICIPATING",
|
| 54 |
+
"143": "SUCCESS",
|
| 55 |
+
"144": "SUCCESS",
|
| 56 |
+
"145": "SUCCESS",
|
| 57 |
+
"146": "SUCCESS",
|
| 58 |
+
"147": "NOT_ALIVE",
|
| 59 |
"148": "SUCCESS",
|
| 60 |
+
"149": "NOT_ALIVE",
|
| 61 |
"15": "SUCCESS",
|
| 62 |
+
"150": "NOT_ALIVE",
|
| 63 |
+
"151": "NOT_ALIVE",
|
| 64 |
+
"152": "NOT_ALIVE",
|
| 65 |
+
"153": "NOT_ALIVE",
|
| 66 |
"154": "NON_PARTICIPATING",
|
| 67 |
+
"155": "SUCCESS",
|
| 68 |
+
"156": "NOT_ALIVE",
|
| 69 |
"157": "NON_PARTICIPATING",
|
| 70 |
"158": "NON_PARTICIPATING",
|
| 71 |
"159": "NON_PARTICIPATING",
|
| 72 |
"16": "SUCCESS",
|
| 73 |
"160": "NON_PARTICIPATING",
|
| 74 |
+
"161": "NOT_ALIVE",
|
| 75 |
+
"162": "NOT_ALIVE",
|
| 76 |
+
"163": "SUCCESS",
|
| 77 |
"164": "NON_PARTICIPATING",
|
| 78 |
"165": "NON_PARTICIPATING",
|
| 79 |
"166": "SUCCESS",
|
| 80 |
+
"167": "NOT_ALIVE",
|
| 81 |
"168": "NON_PARTICIPATING",
|
| 82 |
"169": "NON_PARTICIPATING",
|
| 83 |
+
"17": "NOT_ALIVE",
|
| 84 |
"170": "NON_PARTICIPATING",
|
| 85 |
"171": "NON_PARTICIPATING",
|
| 86 |
+
"172": "NOT_ALIVE",
|
| 87 |
+
"173": "SUCCESS",
|
| 88 |
+
"174": "NOT_ALIVE",
|
| 89 |
+
"175": "NOT_ALIVE",
|
| 90 |
+
"176": "NOT_ALIVE",
|
| 91 |
+
"177": "SUCCESS",
|
| 92 |
+
"178": "NOT_ALIVE",
|
| 93 |
+
"179": "FAIL",
|
| 94 |
+
"18": "NOT_ALIVE",
|
| 95 |
"180": "NON_PARTICIPATING",
|
| 96 |
"181": "SUCCESS",
|
| 97 |
+
"182": "NOT_ALIVE",
|
| 98 |
+
"183": "NOT_ALIVE",
|
| 99 |
+
"184": "NOT_ALIVE",
|
| 100 |
+
"185": "SUCCESS",
|
| 101 |
+
"186": "SUCCESS",
|
| 102 |
+
"187": "SUCCESS",
|
| 103 |
+
"188": "SUCCESS",
|
| 104 |
+
"189": "SUCCESS",
|
| 105 |
+
"19": "NOT_ALIVE",
|
| 106 |
+
"190": "NOT_ALIVE",
|
| 107 |
"191": "NON_PARTICIPATING",
|
| 108 |
+
"192": "NOT_ALIVE",
|
| 109 |
+
"193": "SUCCESS",
|
| 110 |
+
"194": "SUCCESS",
|
| 111 |
+
"195": "SUCCESS",
|
| 112 |
+
"196": "SUCCESS",
|
| 113 |
+
"197": "NOT_ALIVE",
|
| 114 |
"198": "NON_PARTICIPATING",
|
| 115 |
+
"199": "SUCCESS",
|
| 116 |
+
"2": "NOT_ALIVE",
|
| 117 |
+
"20": "NOT_ALIVE",
|
| 118 |
+
"200": "NOT_ALIVE",
|
| 119 |
+
"201": "NOT_ALIVE",
|
| 120 |
"202": "NON_PARTICIPATING",
|
| 121 |
+
"203": "NOT_ALIVE",
|
| 122 |
+
"204": "SUCCESS",
|
| 123 |
+
"205": "SUCCESS",
|
| 124 |
"206": "NON_PARTICIPATING",
|
| 125 |
+
"207": "NOT_ALIVE",
|
| 126 |
+
"208": "SUCCESS",
|
| 127 |
+
"209": "SUCCESS",
|
| 128 |
"21": "SUCCESS",
|
| 129 |
+
"210": "NON_PARTICIPATING",
|
| 130 |
+
"211": "SUCCESS",
|
| 131 |
"212": "NON_PARTICIPATING",
|
| 132 |
"213": "NON_PARTICIPATING",
|
| 133 |
+
"214": "SUCCESS",
|
| 134 |
+
"215": "NOT_ALIVE",
|
| 135 |
+
"216": "SUCCESS",
|
| 136 |
+
"217": "SUCCESS",
|
| 137 |
"218": "NON_PARTICIPATING",
|
| 138 |
+
"219": "SUCCESS",
|
| 139 |
+
"22": "NOT_ALIVE",
|
| 140 |
"220": "NON_PARTICIPATING",
|
| 141 |
"221": "NON_PARTICIPATING",
|
| 142 |
+
"222": "SUCCESS",
|
| 143 |
+
"223": "NOT_ALIVE",
|
| 144 |
"224": "NON_PARTICIPATING",
|
| 145 |
+
"225": "SUCCESS",
|
| 146 |
+
"226": "NOT_ALIVE",
|
| 147 |
+
"227": "NOT_ALIVE",
|
| 148 |
+
"228": "SUCCESS",
|
| 149 |
"229": "NON_PARTICIPATING",
|
| 150 |
+
"23": "NOT_ALIVE",
|
| 151 |
+
"230": "SUCCESS",
|
| 152 |
"231": "NON_PARTICIPATING",
|
| 153 |
+
"232": "NOT_ALIVE",
|
| 154 |
"233": "NON_PARTICIPATING",
|
| 155 |
+
"234": "SUCCESS",
|
| 156 |
+
"235": "NOT_ALIVE",
|
| 157 |
"236": "NON_PARTICIPATING",
|
| 158 |
+
"237": "NOT_ALIVE",
|
| 159 |
+
"238": "SUCCESS",
|
| 160 |
"239": "NON_PARTICIPATING",
|
| 161 |
"24": "SUCCESS",
|
| 162 |
"240": "NON_PARTICIPATING",
|
| 163 |
"241": "NON_PARTICIPATING",
|
| 164 |
"242": "NON_PARTICIPATING",
|
| 165 |
+
"243": "NOT_ALIVE",
|
| 166 |
"244": "NON_PARTICIPATING",
|
| 167 |
+
"245": "NOT_ALIVE",
|
| 168 |
"246": "SUCCESS",
|
| 169 |
"247": "NON_PARTICIPATING",
|
| 170 |
+
"248": "NOT_ALIVE",
|
| 171 |
+
"249": "NOT_ALIVE",
|
| 172 |
+
"25": "SUCCESS",
|
| 173 |
"250": "NON_PARTICIPATING",
|
| 174 |
+
"251": "SUCCESS",
|
| 175 |
+
"252": "NOT_ALIVE",
|
| 176 |
+
"253": "SUCCESS",
|
| 177 |
+
"254": "NOT_ALIVE",
|
| 178 |
"255": "NON_PARTICIPATING",
|
| 179 |
+
"26": "NON_PARTICIPATING",
|
| 180 |
"27": "SUCCESS",
|
| 181 |
"28": "SUCCESS",
|
| 182 |
+
"29": "NOT_ALIVE",
|
| 183 |
"3": "SUCCESS",
|
| 184 |
+
"30": "SUCCESS",
|
| 185 |
+
"31": "NON_PARTICIPATING",
|
| 186 |
+
"32": "NOT_ALIVE",
|
| 187 |
+
"33": "NOT_ALIVE",
|
| 188 |
+
"34": "NON_PARTICIPATING",
|
| 189 |
+
"35": "NOT_ALIVE",
|
| 190 |
+
"36": "NOT_ALIVE",
|
| 191 |
+
"37": "NOT_ALIVE",
|
| 192 |
+
"38": "NOT_ALIVE",
|
| 193 |
+
"39": "NON_PARTICIPATING",
|
| 194 |
+
"4": "NOT_ALIVE",
|
| 195 |
+
"40": "NON_PARTICIPATING",
|
| 196 |
+
"41": "NON_PARTICIPATING",
|
| 197 |
+
"42": "NOT_ALIVE",
|
| 198 |
+
"43": "SUCCESS",
|
| 199 |
"44": "SUCCESS",
|
| 200 |
+
"45": "NOT_ALIVE",
|
| 201 |
"46": "SUCCESS",
|
| 202 |
"47": "SUCCESS",
|
| 203 |
+
"48": "NOT_ALIVE",
|
| 204 |
+
"49": "NON_PARTICIPATING",
|
| 205 |
+
"5": "NOT_ALIVE",
|
| 206 |
+
"50": "NOT_ALIVE",
|
| 207 |
+
"51": "NOT_ALIVE",
|
| 208 |
"52": "SUCCESS",
|
| 209 |
+
"53": "NON_PARTICIPATING",
|
| 210 |
+
"54": "NOT_ALIVE",
|
| 211 |
+
"55": "NON_PARTICIPATING",
|
| 212 |
+
"56": "NOT_ALIVE",
|
| 213 |
+
"57": "SUCCESS",
|
| 214 |
+
"58": "NOT_ALIVE",
|
| 215 |
+
"59": "NOT_ALIVE",
|
| 216 |
"6": "SUCCESS",
|
| 217 |
+
"60": "NOT_ALIVE",
|
| 218 |
+
"61": "NOT_ALIVE",
|
| 219 |
+
"62": "NOT_ALIVE",
|
| 220 |
+
"63": "NOT_ALIVE",
|
| 221 |
"64": "SUCCESS",
|
| 222 |
+
"65": "NOT_ALIVE",
|
| 223 |
+
"66": "NOT_ALIVE",
|
| 224 |
+
"67": "NOT_ALIVE",
|
| 225 |
+
"68": "NON_PARTICIPATING",
|
| 226 |
+
"69": "NON_PARTICIPATING",
|
| 227 |
+
"7": "NOT_ALIVE",
|
| 228 |
+
"70": "NOT_ALIVE",
|
| 229 |
+
"71": "NON_PARTICIPATING",
|
| 230 |
+
"72": "NON_PARTICIPATING",
|
| 231 |
"73": "SUCCESS",
|
| 232 |
+
"74": "NOT_ALIVE",
|
| 233 |
+
"75": "NOT_ALIVE",
|
| 234 |
+
"76": "NOT_ALIVE",
|
| 235 |
"77": "SUCCESS",
|
| 236 |
+
"78": "NOT_ALIVE",
|
| 237 |
+
"79": "NOT_ALIVE",
|
| 238 |
+
"8": "NOT_ALIVE",
|
| 239 |
+
"80": "NOT_ALIVE",
|
| 240 |
+
"81": "NOT_ALIVE",
|
| 241 |
+
"82": "NON_PARTICIPATING",
|
| 242 |
+
"83": "NOT_ALIVE",
|
| 243 |
+
"84": "NOT_ALIVE",
|
| 244 |
+
"85": "NOT_ALIVE",
|
| 245 |
+
"86": "NOT_ALIVE",
|
| 246 |
+
"87": "NOT_ALIVE",
|
| 247 |
"88": "SUCCESS",
|
| 248 |
+
"89": "NON_PARTICIPATING",
|
| 249 |
+
"9": "NOT_ALIVE",
|
| 250 |
+
"90": "NOT_ALIVE",
|
| 251 |
+
"91": "NOT_ALIVE",
|
| 252 |
+
"92": "NOT_ALIVE",
|
| 253 |
+
"93": "NOT_ALIVE",
|
| 254 |
+
"94": "NON_PARTICIPATING",
|
| 255 |
+
"95": "NOT_ALIVE",
|
| 256 |
+
"96": "NOT_ALIVE",
|
| 257 |
+
"97": "NOT_ALIVE",
|
| 258 |
"98": "SUCCESS",
|
| 259 |
+
"99": "NOT_ALIVE"
|
| 260 |
},
|
| 261 |
"architectures": [
|
| 262 |
+
"LlamaForCausalLM"
|
| 263 |
],
|
| 264 |
+
"attention_bias": false,
|
| 265 |
+
"attention_dropout": 0.0,
|
|
|
|
|
|
|
|
|
|
| 266 |
"block_list": [
|
| 267 |
+
6355635,
|
| 268 |
+
6355652
|
| 269 |
],
|
| 270 |
+
"bos_token_id": 1,
|
| 271 |
+
"eos_token_id": 2,
|
| 272 |
+
"hidden_act": "silu",
|
| 273 |
+
"hidden_size": 2048,
|
| 274 |
"initializer_range": 0.02,
|
| 275 |
+
"inner_step": 4,
|
| 276 |
+
"intermediate_size": 5632,
|
| 277 |
+
"last_allreduce_block": 6354618,
|
| 278 |
+
"max_position_embeddings": 2048,
|
| 279 |
+
"mlp_bias": false,
|
| 280 |
+
"model_type": "llama",
|
| 281 |
+
"num_attention_heads": 32,
|
| 282 |
+
"num_hidden_layers": 22,
|
| 283 |
+
"num_key_value_heads": 4,
|
| 284 |
+
"pretraining_tp": 1,
|
| 285 |
+
"rms_norm_eps": 1e-05,
|
| 286 |
+
"rope_scaling": null,
|
| 287 |
+
"rope_theta": 10000.0,
|
| 288 |
+
"tie_word_embeddings": false,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 289 |
"torch_dtype": "float32",
|
| 290 |
"transformers_version": "4.39.3",
|
| 291 |
+
"use_cache": false,
|
| 292 |
+
"vocab_size": 32000
|
| 293 |
}
|
generation_config.json
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"bos_token_id": 1,
|
| 4 |
+
"eos_token_id": 2,
|
| 5 |
+
"transformers_version": "4.39.3",
|
| 6 |
+
"use_cache": false
|
| 7 |
+
}
|
gradients.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1b79ef5325a56da442e26fd385d38c1b27f6444c02411b30d8464e9080570630
|
| 3 |
+
size 88389626
|
inner_optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c611b31267b119151a77fa85da80c3515cdc187c1b7752ce66b20946c1f6533b
|
| 3 |
+
size 4924926906
|
model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:cc3c407c215b5dc27c1accbbb818c068dae4376e1cae22c47ad1144e86c1c31f
|
| 3 |
+
size 4400216536
|