Outer Step 2. Inner Step 5. Batch Size 400
Browse files- config.json +48 -48
- inner_optimizer.pt +1 -1
- model.safetensors +1 -1
config.json
CHANGED
|
@@ -16,14 +16,14 @@
|
|
| 16 |
"108": "NON_PARTICIPATING",
|
| 17 |
"109": "NON_PARTICIPATING",
|
| 18 |
"11": "NON_PARTICIPATING",
|
| 19 |
-
"110": "
|
| 20 |
"111": "NON_PARTICIPATING",
|
| 21 |
"112": "NON_PARTICIPATING",
|
| 22 |
"113": "NON_PARTICIPATING",
|
| 23 |
"114": "NON_PARTICIPATING",
|
| 24 |
-
"115": "
|
| 25 |
"116": "NON_PARTICIPATING",
|
| 26 |
-
"117": "
|
| 27 |
"118": "NON_PARTICIPATING",
|
| 28 |
"119": "NON_PARTICIPATING",
|
| 29 |
"12": "NON_PARTICIPATING",
|
|
@@ -36,9 +36,9 @@
|
|
| 36 |
"126": "SUCCESS",
|
| 37 |
"127": "SUCCESS",
|
| 38 |
"128": "NON_PARTICIPATING",
|
| 39 |
-
"129": "
|
| 40 |
"13": "NON_PARTICIPATING",
|
| 41 |
-
"130": "
|
| 42 |
"131": "NON_PARTICIPATING",
|
| 43 |
"132": "NON_PARTICIPATING",
|
| 44 |
"133": "NON_PARTICIPATING",
|
|
@@ -55,11 +55,11 @@
|
|
| 55 |
"143": "NON_PARTICIPATING",
|
| 56 |
"144": "NON_PARTICIPATING",
|
| 57 |
"145": "SUCCESS",
|
| 58 |
-
"146": "
|
| 59 |
"147": "NON_PARTICIPATING",
|
| 60 |
-
"148": "
|
| 61 |
"149": "NON_PARTICIPATING",
|
| 62 |
-
"15": "
|
| 63 |
"150": "NON_PARTICIPATING",
|
| 64 |
"151": "NON_PARTICIPATING",
|
| 65 |
"152": "NON_PARTICIPATING",
|
|
@@ -75,25 +75,25 @@
|
|
| 75 |
"161": "NON_PARTICIPATING",
|
| 76 |
"162": "NON_PARTICIPATING",
|
| 77 |
"163": "NON_PARTICIPATING",
|
| 78 |
-
"164": "
|
| 79 |
"165": "NON_PARTICIPATING",
|
| 80 |
"166": "SUCCESS",
|
| 81 |
"167": "NON_PARTICIPATING",
|
| 82 |
-
"168": "
|
| 83 |
"169": "NON_PARTICIPATING",
|
| 84 |
"17": "NON_PARTICIPATING",
|
| 85 |
-
"170": "
|
| 86 |
-
"171": "
|
| 87 |
-
"172": "
|
| 88 |
"173": "NON_PARTICIPATING",
|
| 89 |
"174": "SUCCESS",
|
| 90 |
"175": "NON_PARTICIPATING",
|
| 91 |
"176": "NON_PARTICIPATING",
|
| 92 |
"177": "NON_PARTICIPATING",
|
| 93 |
-
"178": "
|
| 94 |
"179": "NON_PARTICIPATING",
|
| 95 |
"18": "NON_PARTICIPATING",
|
| 96 |
-
"180": "
|
| 97 |
"181": "NON_PARTICIPATING",
|
| 98 |
"182": "NON_PARTICIPATING",
|
| 99 |
"183": "NON_PARTICIPATING",
|
|
@@ -104,7 +104,7 @@
|
|
| 104 |
"188": "NON_PARTICIPATING",
|
| 105 |
"189": "SUCCESS",
|
| 106 |
"19": "NON_PARTICIPATING",
|
| 107 |
-
"190": "
|
| 108 |
"191": "NON_PARTICIPATING",
|
| 109 |
"192": "SUCCESS",
|
| 110 |
"193": "NON_PARTICIPATING",
|
|
@@ -116,16 +116,16 @@
|
|
| 116 |
"199": "NON_PARTICIPATING",
|
| 117 |
"2": "NON_PARTICIPATING",
|
| 118 |
"20": "NON_PARTICIPATING",
|
| 119 |
-
"200": "
|
| 120 |
"201": "NON_PARTICIPATING",
|
| 121 |
"202": "NON_PARTICIPATING",
|
| 122 |
"203": "SUCCESS",
|
| 123 |
"204": "NON_PARTICIPATING",
|
| 124 |
-
"205": "
|
| 125 |
"206": "NON_PARTICIPATING",
|
| 126 |
"207": "NON_PARTICIPATING",
|
| 127 |
"208": "NON_PARTICIPATING",
|
| 128 |
-
"209": "
|
| 129 |
"21": "NON_PARTICIPATING",
|
| 130 |
"210": "NON_PARTICIPATING",
|
| 131 |
"211": "SUCCESS",
|
|
@@ -137,33 +137,33 @@
|
|
| 137 |
"217": "NON_PARTICIPATING",
|
| 138 |
"218": "NON_PARTICIPATING",
|
| 139 |
"219": "NON_PARTICIPATING",
|
| 140 |
-
"22": "
|
| 141 |
"220": "NON_PARTICIPATING",
|
| 142 |
"221": "NON_PARTICIPATING",
|
| 143 |
"222": "NON_PARTICIPATING",
|
| 144 |
"223": "NON_PARTICIPATING",
|
| 145 |
"224": "NON_PARTICIPATING",
|
| 146 |
"225": "NON_PARTICIPATING",
|
| 147 |
-
"226": "
|
| 148 |
"227": "NON_PARTICIPATING",
|
| 149 |
"228": "NON_PARTICIPATING",
|
| 150 |
"229": "NON_PARTICIPATING",
|
| 151 |
"23": "NON_PARTICIPATING",
|
| 152 |
-
"230": "
|
| 153 |
"231": "NON_PARTICIPATING",
|
| 154 |
"232": "NON_PARTICIPATING",
|
| 155 |
"233": "SUCCESS",
|
| 156 |
-
"234": "
|
| 157 |
"235": "SUCCESS",
|
| 158 |
"236": "NON_PARTICIPATING",
|
| 159 |
-
"237": "
|
| 160 |
-
"238": "
|
| 161 |
-
"239": "
|
| 162 |
"24": "NON_PARTICIPATING",
|
| 163 |
"240": "SUCCESS",
|
| 164 |
"241": "NON_PARTICIPATING",
|
| 165 |
"242": "NON_PARTICIPATING",
|
| 166 |
-
"243": "
|
| 167 |
"244": "SUCCESS",
|
| 168 |
"245": "NON_PARTICIPATING",
|
| 169 |
"246": "NON_PARTICIPATING",
|
|
@@ -173,14 +173,14 @@
|
|
| 173 |
"25": "SUCCESS",
|
| 174 |
"250": "NON_PARTICIPATING",
|
| 175 |
"251": "SUCCESS",
|
| 176 |
-
"252": "
|
| 177 |
"253": "NON_PARTICIPATING",
|
| 178 |
"254": "NON_PARTICIPATING",
|
| 179 |
"255": "NON_PARTICIPATING",
|
| 180 |
"26": "NON_PARTICIPATING",
|
| 181 |
"27": "NON_PARTICIPATING",
|
| 182 |
"28": "NON_PARTICIPATING",
|
| 183 |
-
"29": "
|
| 184 |
"3": "NON_PARTICIPATING",
|
| 185 |
"30": "NON_PARTICIPATING",
|
| 186 |
"31": "NON_PARTICIPATING",
|
|
@@ -200,12 +200,12 @@
|
|
| 200 |
"44": "NON_PARTICIPATING",
|
| 201 |
"45": "NON_PARTICIPATING",
|
| 202 |
"46": "NON_PARTICIPATING",
|
| 203 |
-
"47": "
|
| 204 |
"48": "NON_PARTICIPATING",
|
| 205 |
"49": "NON_PARTICIPATING",
|
| 206 |
"5": "NON_PARTICIPATING",
|
| 207 |
"50": "NON_PARTICIPATING",
|
| 208 |
-
"51": "
|
| 209 |
"52": "NON_PARTICIPATING",
|
| 210 |
"53": "NON_PARTICIPATING",
|
| 211 |
"54": "NON_PARTICIPATING",
|
|
@@ -214,7 +214,7 @@
|
|
| 214 |
"57": "NON_PARTICIPATING",
|
| 215 |
"58": "SUCCESS",
|
| 216 |
"59": "NON_PARTICIPATING",
|
| 217 |
-
"6": "
|
| 218 |
"60": "NON_PARTICIPATING",
|
| 219 |
"61": "NON_PARTICIPATING",
|
| 220 |
"62": "SUCCESS",
|
|
@@ -230,8 +230,8 @@
|
|
| 230 |
"71": "NON_PARTICIPATING",
|
| 231 |
"72": "NON_PARTICIPATING",
|
| 232 |
"73": "NON_PARTICIPATING",
|
| 233 |
-
"74": "
|
| 234 |
-
"75": "
|
| 235 |
"76": "NON_PARTICIPATING",
|
| 236 |
"77": "NON_PARTICIPATING",
|
| 237 |
"78": "NON_PARTICIPATING",
|
|
@@ -241,7 +241,7 @@
|
|
| 241 |
"81": "NON_PARTICIPATING",
|
| 242 |
"82": "NON_PARTICIPATING",
|
| 243 |
"83": "SUCCESS",
|
| 244 |
-
"84": "
|
| 245 |
"85": "NON_PARTICIPATING",
|
| 246 |
"86": "SUCCESS",
|
| 247 |
"87": "NON_PARTICIPATING",
|
|
@@ -252,7 +252,7 @@
|
|
| 252 |
"91": "NON_PARTICIPATING",
|
| 253 |
"92": "NON_PARTICIPATING",
|
| 254 |
"93": "NON_PARTICIPATING",
|
| 255 |
-
"94": "
|
| 256 |
"95": "SUCCESS",
|
| 257 |
"96": "NON_PARTICIPATING",
|
| 258 |
"97": "SUCCESS",
|
|
@@ -268,25 +268,25 @@
|
|
| 268 |
"AutoModelForCausalLM": "distributed/optimized-gpt2-500m--modeling_gpt_optimized.GPTOptim"
|
| 269 |
},
|
| 270 |
"block_list": [
|
| 271 |
-
|
| 272 |
-
|
| 273 |
-
|
| 274 |
-
|
| 275 |
-
|
| 276 |
-
|
| 277 |
-
|
| 278 |
-
|
| 279 |
-
|
| 280 |
-
|
| 281 |
],
|
| 282 |
"block_size": 1024,
|
| 283 |
"bos_token_id": 50256,
|
| 284 |
"embd_pdrop": 0.1,
|
| 285 |
"eos_token_id": 50256,
|
| 286 |
"initializer_range": 0.02,
|
| 287 |
-
"inner_step":
|
| 288 |
"inner_steps": 0,
|
| 289 |
-
"last_allreduce_block":
|
| 290 |
"layer_norm_epsilon": 1e-05,
|
| 291 |
"model_type": "gpt_optimized",
|
| 292 |
"n_embd": 1280,
|
|
|
|
| 16 |
"108": "NON_PARTICIPATING",
|
| 17 |
"109": "NON_PARTICIPATING",
|
| 18 |
"11": "NON_PARTICIPATING",
|
| 19 |
+
"110": "FAIL",
|
| 20 |
"111": "NON_PARTICIPATING",
|
| 21 |
"112": "NON_PARTICIPATING",
|
| 22 |
"113": "NON_PARTICIPATING",
|
| 23 |
"114": "NON_PARTICIPATING",
|
| 24 |
+
"115": "FAIL",
|
| 25 |
"116": "NON_PARTICIPATING",
|
| 26 |
+
"117": "FAIL",
|
| 27 |
"118": "NON_PARTICIPATING",
|
| 28 |
"119": "NON_PARTICIPATING",
|
| 29 |
"12": "NON_PARTICIPATING",
|
|
|
|
| 36 |
"126": "SUCCESS",
|
| 37 |
"127": "SUCCESS",
|
| 38 |
"128": "NON_PARTICIPATING",
|
| 39 |
+
"129": "FAIL",
|
| 40 |
"13": "NON_PARTICIPATING",
|
| 41 |
+
"130": "SUCCESS",
|
| 42 |
"131": "NON_PARTICIPATING",
|
| 43 |
"132": "NON_PARTICIPATING",
|
| 44 |
"133": "NON_PARTICIPATING",
|
|
|
|
| 55 |
"143": "NON_PARTICIPATING",
|
| 56 |
"144": "NON_PARTICIPATING",
|
| 57 |
"145": "SUCCESS",
|
| 58 |
+
"146": "FAIL",
|
| 59 |
"147": "NON_PARTICIPATING",
|
| 60 |
+
"148": "FAIL",
|
| 61 |
"149": "NON_PARTICIPATING",
|
| 62 |
+
"15": "FAIL",
|
| 63 |
"150": "NON_PARTICIPATING",
|
| 64 |
"151": "NON_PARTICIPATING",
|
| 65 |
"152": "NON_PARTICIPATING",
|
|
|
|
| 75 |
"161": "NON_PARTICIPATING",
|
| 76 |
"162": "NON_PARTICIPATING",
|
| 77 |
"163": "NON_PARTICIPATING",
|
| 78 |
+
"164": "FAIL",
|
| 79 |
"165": "NON_PARTICIPATING",
|
| 80 |
"166": "SUCCESS",
|
| 81 |
"167": "NON_PARTICIPATING",
|
| 82 |
+
"168": "SUCCESS",
|
| 83 |
"169": "NON_PARTICIPATING",
|
| 84 |
"17": "NON_PARTICIPATING",
|
| 85 |
+
"170": "FAIL",
|
| 86 |
+
"171": "SUCCESS",
|
| 87 |
+
"172": "NON_PARTICIPATING",
|
| 88 |
"173": "NON_PARTICIPATING",
|
| 89 |
"174": "SUCCESS",
|
| 90 |
"175": "NON_PARTICIPATING",
|
| 91 |
"176": "NON_PARTICIPATING",
|
| 92 |
"177": "NON_PARTICIPATING",
|
| 93 |
+
"178": "FAIL",
|
| 94 |
"179": "NON_PARTICIPATING",
|
| 95 |
"18": "NON_PARTICIPATING",
|
| 96 |
+
"180": "FAIL",
|
| 97 |
"181": "NON_PARTICIPATING",
|
| 98 |
"182": "NON_PARTICIPATING",
|
| 99 |
"183": "NON_PARTICIPATING",
|
|
|
|
| 104 |
"188": "NON_PARTICIPATING",
|
| 105 |
"189": "SUCCESS",
|
| 106 |
"19": "NON_PARTICIPATING",
|
| 107 |
+
"190": "FAIL",
|
| 108 |
"191": "NON_PARTICIPATING",
|
| 109 |
"192": "SUCCESS",
|
| 110 |
"193": "NON_PARTICIPATING",
|
|
|
|
| 116 |
"199": "NON_PARTICIPATING",
|
| 117 |
"2": "NON_PARTICIPATING",
|
| 118 |
"20": "NON_PARTICIPATING",
|
| 119 |
+
"200": "FAIL",
|
| 120 |
"201": "NON_PARTICIPATING",
|
| 121 |
"202": "NON_PARTICIPATING",
|
| 122 |
"203": "SUCCESS",
|
| 123 |
"204": "NON_PARTICIPATING",
|
| 124 |
+
"205": "FAIL",
|
| 125 |
"206": "NON_PARTICIPATING",
|
| 126 |
"207": "NON_PARTICIPATING",
|
| 127 |
"208": "NON_PARTICIPATING",
|
| 128 |
+
"209": "FAIL",
|
| 129 |
"21": "NON_PARTICIPATING",
|
| 130 |
"210": "NON_PARTICIPATING",
|
| 131 |
"211": "SUCCESS",
|
|
|
|
| 137 |
"217": "NON_PARTICIPATING",
|
| 138 |
"218": "NON_PARTICIPATING",
|
| 139 |
"219": "NON_PARTICIPATING",
|
| 140 |
+
"22": "FAIL",
|
| 141 |
"220": "NON_PARTICIPATING",
|
| 142 |
"221": "NON_PARTICIPATING",
|
| 143 |
"222": "NON_PARTICIPATING",
|
| 144 |
"223": "NON_PARTICIPATING",
|
| 145 |
"224": "NON_PARTICIPATING",
|
| 146 |
"225": "NON_PARTICIPATING",
|
| 147 |
+
"226": "SUCCESS",
|
| 148 |
"227": "NON_PARTICIPATING",
|
| 149 |
"228": "NON_PARTICIPATING",
|
| 150 |
"229": "NON_PARTICIPATING",
|
| 151 |
"23": "NON_PARTICIPATING",
|
| 152 |
+
"230": "FAIL",
|
| 153 |
"231": "NON_PARTICIPATING",
|
| 154 |
"232": "NON_PARTICIPATING",
|
| 155 |
"233": "SUCCESS",
|
| 156 |
+
"234": "NON_PARTICIPATING",
|
| 157 |
"235": "SUCCESS",
|
| 158 |
"236": "NON_PARTICIPATING",
|
| 159 |
+
"237": "FAIL",
|
| 160 |
+
"238": "FAIL",
|
| 161 |
+
"239": "FAIL",
|
| 162 |
"24": "NON_PARTICIPATING",
|
| 163 |
"240": "SUCCESS",
|
| 164 |
"241": "NON_PARTICIPATING",
|
| 165 |
"242": "NON_PARTICIPATING",
|
| 166 |
+
"243": "FAIL",
|
| 167 |
"244": "SUCCESS",
|
| 168 |
"245": "NON_PARTICIPATING",
|
| 169 |
"246": "NON_PARTICIPATING",
|
|
|
|
| 173 |
"25": "SUCCESS",
|
| 174 |
"250": "NON_PARTICIPATING",
|
| 175 |
"251": "SUCCESS",
|
| 176 |
+
"252": "FAIL",
|
| 177 |
"253": "NON_PARTICIPATING",
|
| 178 |
"254": "NON_PARTICIPATING",
|
| 179 |
"255": "NON_PARTICIPATING",
|
| 180 |
"26": "NON_PARTICIPATING",
|
| 181 |
"27": "NON_PARTICIPATING",
|
| 182 |
"28": "NON_PARTICIPATING",
|
| 183 |
+
"29": "NON_PARTICIPATING",
|
| 184 |
"3": "NON_PARTICIPATING",
|
| 185 |
"30": "NON_PARTICIPATING",
|
| 186 |
"31": "NON_PARTICIPATING",
|
|
|
|
| 200 |
"44": "NON_PARTICIPATING",
|
| 201 |
"45": "NON_PARTICIPATING",
|
| 202 |
"46": "NON_PARTICIPATING",
|
| 203 |
+
"47": "FAIL",
|
| 204 |
"48": "NON_PARTICIPATING",
|
| 205 |
"49": "NON_PARTICIPATING",
|
| 206 |
"5": "NON_PARTICIPATING",
|
| 207 |
"50": "NON_PARTICIPATING",
|
| 208 |
+
"51": "SUCCESS",
|
| 209 |
"52": "NON_PARTICIPATING",
|
| 210 |
"53": "NON_PARTICIPATING",
|
| 211 |
"54": "NON_PARTICIPATING",
|
|
|
|
| 214 |
"57": "NON_PARTICIPATING",
|
| 215 |
"58": "SUCCESS",
|
| 216 |
"59": "NON_PARTICIPATING",
|
| 217 |
+
"6": "FAIL",
|
| 218 |
"60": "NON_PARTICIPATING",
|
| 219 |
"61": "NON_PARTICIPATING",
|
| 220 |
"62": "SUCCESS",
|
|
|
|
| 230 |
"71": "NON_PARTICIPATING",
|
| 231 |
"72": "NON_PARTICIPATING",
|
| 232 |
"73": "NON_PARTICIPATING",
|
| 233 |
+
"74": "FAIL",
|
| 234 |
+
"75": "FAIL",
|
| 235 |
"76": "NON_PARTICIPATING",
|
| 236 |
"77": "NON_PARTICIPATING",
|
| 237 |
"78": "NON_PARTICIPATING",
|
|
|
|
| 241 |
"81": "NON_PARTICIPATING",
|
| 242 |
"82": "NON_PARTICIPATING",
|
| 243 |
"83": "SUCCESS",
|
| 244 |
+
"84": "FAIL",
|
| 245 |
"85": "NON_PARTICIPATING",
|
| 246 |
"86": "SUCCESS",
|
| 247 |
"87": "NON_PARTICIPATING",
|
|
|
|
| 252 |
"91": "NON_PARTICIPATING",
|
| 253 |
"92": "NON_PARTICIPATING",
|
| 254 |
"93": "NON_PARTICIPATING",
|
| 255 |
+
"94": "FAIL",
|
| 256 |
"95": "SUCCESS",
|
| 257 |
"96": "NON_PARTICIPATING",
|
| 258 |
"97": "SUCCESS",
|
|
|
|
| 268 |
"AutoModelForCausalLM": "distributed/optimized-gpt2-500m--modeling_gpt_optimized.GPTOptim"
|
| 269 |
},
|
| 270 |
"block_list": [
|
| 271 |
+
5146894,
|
| 272 |
+
5146899,
|
| 273 |
+
5146903,
|
| 274 |
+
5146908,
|
| 275 |
+
5146914,
|
| 276 |
+
5146918,
|
| 277 |
+
5146923,
|
| 278 |
+
5146929,
|
| 279 |
+
5146933,
|
| 280 |
+
5146938
|
| 281 |
],
|
| 282 |
"block_size": 1024,
|
| 283 |
"bos_token_id": 50256,
|
| 284 |
"embd_pdrop": 0.1,
|
| 285 |
"eos_token_id": 50256,
|
| 286 |
"initializer_range": 0.02,
|
| 287 |
+
"inner_step": 5,
|
| 288 |
"inner_steps": 0,
|
| 289 |
+
"last_allreduce_block": 5146835,
|
| 290 |
"layer_norm_epsilon": 1e-05,
|
| 291 |
"model_type": "gpt_optimized",
|
| 292 |
"n_embd": 1280,
|
inner_optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 8081781770
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:acc55f508f416f1bbbbe25dc9e8fc57dbfdba9c40b3de72a36591719c741adbf
|
| 3 |
size 8081781770
|
model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 4040701744
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:fbd754c823fda1915f654206c022efa5b3cd540b6246a1be74b68f44c0ba7a36
|
| 3 |
size 4040701744
|