Run 4. Outer Step 25. Inner Step 0.
Browse files- config.json +39 -45
- inner_optimizer.pt +1 -1
- model.safetensors +1 -1
config.json
CHANGED
|
@@ -7,34 +7,34 @@
|
|
| 7 |
"10": "SUCCESS",
|
| 8 |
"100": "NON_PARTICIPATING",
|
| 9 |
"101": "SUCCESS",
|
| 10 |
-
"102": "
|
| 11 |
"103": "NON_PARTICIPATING",
|
| 12 |
"104": "NON_PARTICIPATING",
|
| 13 |
"105": "NON_PARTICIPATING",
|
| 14 |
"106": "SUCCESS",
|
| 15 |
-
"107": "
|
| 16 |
"108": "NON_PARTICIPATING",
|
| 17 |
"109": "NON_PARTICIPATING",
|
| 18 |
"11": "NON_PARTICIPATING",
|
| 19 |
-
"110": "
|
| 20 |
"111": "SUCCESS",
|
| 21 |
"112": "NON_PARTICIPATING",
|
| 22 |
"113": "SUCCESS",
|
| 23 |
-
"114": "
|
| 24 |
"115": "NON_PARTICIPATING",
|
| 25 |
"116": "NON_PARTICIPATING",
|
| 26 |
"117": "NON_PARTICIPATING",
|
| 27 |
-
"118": "
|
| 28 |
"119": "NON_PARTICIPATING",
|
| 29 |
"12": "SUCCESS",
|
| 30 |
"120": "SUCCESS",
|
| 31 |
"121": "SUCCESS",
|
| 32 |
-
"122": "
|
| 33 |
-
"123": "
|
| 34 |
"124": "NON_PARTICIPATING",
|
| 35 |
"125": "NON_PARTICIPATING",
|
| 36 |
"126": "NON_PARTICIPATING",
|
| 37 |
-
"127": "
|
| 38 |
"128": "NON_PARTICIPATING",
|
| 39 |
"129": "NON_PARTICIPATING",
|
| 40 |
"13": "NON_PARTICIPATING",
|
|
@@ -46,17 +46,17 @@
|
|
| 46 |
"135": "NON_PARTICIPATING",
|
| 47 |
"136": "NON_PARTICIPATING",
|
| 48 |
"137": "NON_PARTICIPATING",
|
| 49 |
-
"138": "
|
| 50 |
"139": "NON_PARTICIPATING",
|
| 51 |
"14": "NON_PARTICIPATING",
|
| 52 |
"140": "NON_PARTICIPATING",
|
| 53 |
"141": "NON_PARTICIPATING",
|
| 54 |
"142": "NON_PARTICIPATING",
|
| 55 |
-
"143": "
|
| 56 |
"144": "NON_PARTICIPATING",
|
| 57 |
"145": "SUCCESS",
|
| 58 |
"146": "NON_PARTICIPATING",
|
| 59 |
-
"147": "
|
| 60 |
"148": "SUCCESS",
|
| 61 |
"149": "NON_PARTICIPATING",
|
| 62 |
"15": "SUCCESS",
|
|
@@ -64,12 +64,12 @@
|
|
| 64 |
"151": "NON_PARTICIPATING",
|
| 65 |
"152": "NON_PARTICIPATING",
|
| 66 |
"153": "NON_PARTICIPATING",
|
| 67 |
-
"154": "
|
| 68 |
"155": "NON_PARTICIPATING",
|
| 69 |
-
"156": "
|
| 70 |
"157": "NON_PARTICIPATING",
|
| 71 |
"158": "NON_PARTICIPATING",
|
| 72 |
-
"159": "
|
| 73 |
"16": "SUCCESS",
|
| 74 |
"160": "NON_PARTICIPATING",
|
| 75 |
"161": "NON_PARTICIPATING",
|
|
@@ -79,18 +79,18 @@
|
|
| 79 |
"165": "NON_PARTICIPATING",
|
| 80 |
"166": "NON_PARTICIPATING",
|
| 81 |
"167": "NON_PARTICIPATING",
|
| 82 |
-
"168": "
|
| 83 |
"169": "NON_PARTICIPATING",
|
| 84 |
"17": "SUCCESS",
|
| 85 |
"170": "SUCCESS",
|
| 86 |
-
"171": "
|
| 87 |
"172": "NON_PARTICIPATING",
|
| 88 |
"173": "NON_PARTICIPATING",
|
| 89 |
"174": "NON_PARTICIPATING",
|
| 90 |
"175": "SUCCESS",
|
| 91 |
"176": "NON_PARTICIPATING",
|
| 92 |
"177": "SUCCESS",
|
| 93 |
-
"178": "
|
| 94 |
"179": "NON_PARTICIPATING",
|
| 95 |
"18": "SUCCESS",
|
| 96 |
"180": "NON_PARTICIPATING",
|
|
@@ -99,7 +99,7 @@
|
|
| 99 |
"183": "NON_PARTICIPATING",
|
| 100 |
"184": "NON_PARTICIPATING",
|
| 101 |
"185": "NON_PARTICIPATING",
|
| 102 |
-
"186": "
|
| 103 |
"187": "SUCCESS",
|
| 104 |
"188": "NON_PARTICIPATING",
|
| 105 |
"189": "NON_PARTICIPATING",
|
|
@@ -109,14 +109,14 @@
|
|
| 109 |
"192": "NON_PARTICIPATING",
|
| 110 |
"193": "NON_PARTICIPATING",
|
| 111 |
"194": "NON_PARTICIPATING",
|
| 112 |
-
"195": "
|
| 113 |
"196": "NON_PARTICIPATING",
|
| 114 |
"197": "NON_PARTICIPATING",
|
| 115 |
"198": "NON_PARTICIPATING",
|
| 116 |
"199": "NON_PARTICIPATING",
|
| 117 |
"2": "SUCCESS",
|
| 118 |
"20": "SUCCESS",
|
| 119 |
-
"200": "
|
| 120 |
"201": "NON_PARTICIPATING",
|
| 121 |
"202": "NON_PARTICIPATING",
|
| 122 |
"203": "SUCCESS",
|
|
@@ -128,14 +128,14 @@
|
|
| 128 |
"209": "NON_PARTICIPATING",
|
| 129 |
"21": "SUCCESS",
|
| 130 |
"210": "NON_PARTICIPATING",
|
| 131 |
-
"211": "
|
| 132 |
"212": "NON_PARTICIPATING",
|
| 133 |
"213": "SUCCESS",
|
| 134 |
-
"214": "
|
| 135 |
"215": "NON_PARTICIPATING",
|
| 136 |
"216": "NON_PARTICIPATING",
|
| 137 |
"217": "NON_PARTICIPATING",
|
| 138 |
-
"218": "
|
| 139 |
"219": "NON_PARTICIPATING",
|
| 140 |
"22": "SUCCESS",
|
| 141 |
"220": "NON_PARTICIPATING",
|
|
@@ -148,8 +148,8 @@
|
|
| 148 |
"227": "NON_PARTICIPATING",
|
| 149 |
"228": "NON_PARTICIPATING",
|
| 150 |
"229": "NON_PARTICIPATING",
|
| 151 |
-
"23": "
|
| 152 |
-
"230": "
|
| 153 |
"231": "NON_PARTICIPATING",
|
| 154 |
"232": "NON_PARTICIPATING",
|
| 155 |
"233": "NON_PARTICIPATING",
|
|
@@ -164,7 +164,7 @@
|
|
| 164 |
"241": "SUCCESS",
|
| 165 |
"242": "NON_PARTICIPATING",
|
| 166 |
"243": "NON_PARTICIPATING",
|
| 167 |
-
"244": "
|
| 168 |
"245": "NON_PARTICIPATING",
|
| 169 |
"246": "NON_PARTICIPATING",
|
| 170 |
"247": "NON_PARTICIPATING",
|
|
@@ -175,11 +175,11 @@
|
|
| 175 |
"251": "NON_PARTICIPATING",
|
| 176 |
"252": "NON_PARTICIPATING",
|
| 177 |
"253": "NON_PARTICIPATING",
|
| 178 |
-
"254": "
|
| 179 |
"255": "NON_PARTICIPATING",
|
| 180 |
"26": "SUCCESS",
|
| 181 |
"27": "SUCCESS",
|
| 182 |
-
"28": "
|
| 183 |
"29": "SUCCESS",
|
| 184 |
"3": "SUCCESS",
|
| 185 |
"30": "NON_PARTICIPATING",
|
|
@@ -191,7 +191,7 @@
|
|
| 191 |
"36": "NON_PARTICIPATING",
|
| 192 |
"37": "SUCCESS",
|
| 193 |
"38": "NON_PARTICIPATING",
|
| 194 |
-
"39": "
|
| 195 |
"4": "NON_PARTICIPATING",
|
| 196 |
"40": "SUCCESS",
|
| 197 |
"41": "SUCCESS",
|
|
@@ -223,7 +223,7 @@
|
|
| 223 |
"65": "SUCCESS",
|
| 224 |
"66": "NON_PARTICIPATING",
|
| 225 |
"67": "SUCCESS",
|
| 226 |
-
"68": "
|
| 227 |
"69": "NON_PARTICIPATING",
|
| 228 |
"7": "SUCCESS",
|
| 229 |
"70": "NON_PARTICIPATING",
|
|
@@ -240,7 +240,7 @@
|
|
| 240 |
"80": "SUCCESS",
|
| 241 |
"81": "NON_PARTICIPATING",
|
| 242 |
"82": "SUCCESS",
|
| 243 |
-
"83": "
|
| 244 |
"84": "SUCCESS",
|
| 245 |
"85": "NON_PARTICIPATING",
|
| 246 |
"86": "NON_PARTICIPATING",
|
|
@@ -250,7 +250,7 @@
|
|
| 250 |
"9": "NON_PARTICIPATING",
|
| 251 |
"90": "SUCCESS",
|
| 252 |
"91": "NON_PARTICIPATING",
|
| 253 |
-
"92": "
|
| 254 |
"93": "NON_PARTICIPATING",
|
| 255 |
"94": "SUCCESS",
|
| 256 |
"95": "SUCCESS",
|
|
@@ -268,26 +268,20 @@
|
|
| 268 |
"AutoModelForCausalLM": "distributed/optimized-gpt2-500m--modeling_gpt_optimized.GPTOptim"
|
| 269 |
},
|
| 270 |
"block_list": [
|
| 271 |
-
|
| 272 |
-
|
| 273 |
-
|
| 274 |
-
|
| 275 |
-
|
| 276 |
-
5533389,
|
| 277 |
-
5533392,
|
| 278 |
-
5533395,
|
| 279 |
-
5533397,
|
| 280 |
-
5533399,
|
| 281 |
-
5533402
|
| 282 |
],
|
| 283 |
"block_size": 1024,
|
| 284 |
"bos_token_id": 50256,
|
| 285 |
"embd_pdrop": 0.1,
|
| 286 |
"eos_token_id": 50256,
|
| 287 |
"initializer_range": 0.02,
|
| 288 |
-
"inner_step":
|
| 289 |
"inner_steps": 0,
|
| 290 |
-
"last_allreduce_block":
|
| 291 |
"layer_norm_epsilon": 1e-05,
|
| 292 |
"model_type": "gpt_optimized",
|
| 293 |
"n_embd": 1280,
|
|
|
|
| 7 |
"10": "SUCCESS",
|
| 8 |
"100": "NON_PARTICIPATING",
|
| 9 |
"101": "SUCCESS",
|
| 10 |
+
"102": "NON_PARTICIPATING",
|
| 11 |
"103": "NON_PARTICIPATING",
|
| 12 |
"104": "NON_PARTICIPATING",
|
| 13 |
"105": "NON_PARTICIPATING",
|
| 14 |
"106": "SUCCESS",
|
| 15 |
+
"107": "NON_PARTICIPATING",
|
| 16 |
"108": "NON_PARTICIPATING",
|
| 17 |
"109": "NON_PARTICIPATING",
|
| 18 |
"11": "NON_PARTICIPATING",
|
| 19 |
+
"110": "NON_PARTICIPATING",
|
| 20 |
"111": "SUCCESS",
|
| 21 |
"112": "NON_PARTICIPATING",
|
| 22 |
"113": "SUCCESS",
|
| 23 |
+
"114": "NON_PARTICIPATING",
|
| 24 |
"115": "NON_PARTICIPATING",
|
| 25 |
"116": "NON_PARTICIPATING",
|
| 26 |
"117": "NON_PARTICIPATING",
|
| 27 |
+
"118": "NON_PARTICIPATING",
|
| 28 |
"119": "NON_PARTICIPATING",
|
| 29 |
"12": "SUCCESS",
|
| 30 |
"120": "SUCCESS",
|
| 31 |
"121": "SUCCESS",
|
| 32 |
+
"122": "SUCCESS",
|
| 33 |
+
"123": "NON_PARTICIPATING",
|
| 34 |
"124": "NON_PARTICIPATING",
|
| 35 |
"125": "NON_PARTICIPATING",
|
| 36 |
"126": "NON_PARTICIPATING",
|
| 37 |
+
"127": "NON_PARTICIPATING",
|
| 38 |
"128": "NON_PARTICIPATING",
|
| 39 |
"129": "NON_PARTICIPATING",
|
| 40 |
"13": "NON_PARTICIPATING",
|
|
|
|
| 46 |
"135": "NON_PARTICIPATING",
|
| 47 |
"136": "NON_PARTICIPATING",
|
| 48 |
"137": "NON_PARTICIPATING",
|
| 49 |
+
"138": "SUCCESS",
|
| 50 |
"139": "NON_PARTICIPATING",
|
| 51 |
"14": "NON_PARTICIPATING",
|
| 52 |
"140": "NON_PARTICIPATING",
|
| 53 |
"141": "NON_PARTICIPATING",
|
| 54 |
"142": "NON_PARTICIPATING",
|
| 55 |
+
"143": "NON_PARTICIPATING",
|
| 56 |
"144": "NON_PARTICIPATING",
|
| 57 |
"145": "SUCCESS",
|
| 58 |
"146": "NON_PARTICIPATING",
|
| 59 |
+
"147": "NON_PARTICIPATING",
|
| 60 |
"148": "SUCCESS",
|
| 61 |
"149": "NON_PARTICIPATING",
|
| 62 |
"15": "SUCCESS",
|
|
|
|
| 64 |
"151": "NON_PARTICIPATING",
|
| 65 |
"152": "NON_PARTICIPATING",
|
| 66 |
"153": "NON_PARTICIPATING",
|
| 67 |
+
"154": "SUCCESS",
|
| 68 |
"155": "NON_PARTICIPATING",
|
| 69 |
+
"156": "NON_PARTICIPATING",
|
| 70 |
"157": "NON_PARTICIPATING",
|
| 71 |
"158": "NON_PARTICIPATING",
|
| 72 |
+
"159": "SUCCESS",
|
| 73 |
"16": "SUCCESS",
|
| 74 |
"160": "NON_PARTICIPATING",
|
| 75 |
"161": "NON_PARTICIPATING",
|
|
|
|
| 79 |
"165": "NON_PARTICIPATING",
|
| 80 |
"166": "NON_PARTICIPATING",
|
| 81 |
"167": "NON_PARTICIPATING",
|
| 82 |
+
"168": "NON_PARTICIPATING",
|
| 83 |
"169": "NON_PARTICIPATING",
|
| 84 |
"17": "SUCCESS",
|
| 85 |
"170": "SUCCESS",
|
| 86 |
+
"171": "NON_PARTICIPATING",
|
| 87 |
"172": "NON_PARTICIPATING",
|
| 88 |
"173": "NON_PARTICIPATING",
|
| 89 |
"174": "NON_PARTICIPATING",
|
| 90 |
"175": "SUCCESS",
|
| 91 |
"176": "NON_PARTICIPATING",
|
| 92 |
"177": "SUCCESS",
|
| 93 |
+
"178": "SUCCESS",
|
| 94 |
"179": "NON_PARTICIPATING",
|
| 95 |
"18": "SUCCESS",
|
| 96 |
"180": "NON_PARTICIPATING",
|
|
|
|
| 99 |
"183": "NON_PARTICIPATING",
|
| 100 |
"184": "NON_PARTICIPATING",
|
| 101 |
"185": "NON_PARTICIPATING",
|
| 102 |
+
"186": "NON_PARTICIPATING",
|
| 103 |
"187": "SUCCESS",
|
| 104 |
"188": "NON_PARTICIPATING",
|
| 105 |
"189": "NON_PARTICIPATING",
|
|
|
|
| 109 |
"192": "NON_PARTICIPATING",
|
| 110 |
"193": "NON_PARTICIPATING",
|
| 111 |
"194": "NON_PARTICIPATING",
|
| 112 |
+
"195": "SUCCESS",
|
| 113 |
"196": "NON_PARTICIPATING",
|
| 114 |
"197": "NON_PARTICIPATING",
|
| 115 |
"198": "NON_PARTICIPATING",
|
| 116 |
"199": "NON_PARTICIPATING",
|
| 117 |
"2": "SUCCESS",
|
| 118 |
"20": "SUCCESS",
|
| 119 |
+
"200": "NON_PARTICIPATING",
|
| 120 |
"201": "NON_PARTICIPATING",
|
| 121 |
"202": "NON_PARTICIPATING",
|
| 122 |
"203": "SUCCESS",
|
|
|
|
| 128 |
"209": "NON_PARTICIPATING",
|
| 129 |
"21": "SUCCESS",
|
| 130 |
"210": "NON_PARTICIPATING",
|
| 131 |
+
"211": "SUCCESS",
|
| 132 |
"212": "NON_PARTICIPATING",
|
| 133 |
"213": "SUCCESS",
|
| 134 |
+
"214": "NON_PARTICIPATING",
|
| 135 |
"215": "NON_PARTICIPATING",
|
| 136 |
"216": "NON_PARTICIPATING",
|
| 137 |
"217": "NON_PARTICIPATING",
|
| 138 |
+
"218": "SUCCESS",
|
| 139 |
"219": "NON_PARTICIPATING",
|
| 140 |
"22": "SUCCESS",
|
| 141 |
"220": "NON_PARTICIPATING",
|
|
|
|
| 148 |
"227": "NON_PARTICIPATING",
|
| 149 |
"228": "NON_PARTICIPATING",
|
| 150 |
"229": "NON_PARTICIPATING",
|
| 151 |
+
"23": "NON_PARTICIPATING",
|
| 152 |
+
"230": "NON_PARTICIPATING",
|
| 153 |
"231": "NON_PARTICIPATING",
|
| 154 |
"232": "NON_PARTICIPATING",
|
| 155 |
"233": "NON_PARTICIPATING",
|
|
|
|
| 164 |
"241": "SUCCESS",
|
| 165 |
"242": "NON_PARTICIPATING",
|
| 166 |
"243": "NON_PARTICIPATING",
|
| 167 |
+
"244": "NON_PARTICIPATING",
|
| 168 |
"245": "NON_PARTICIPATING",
|
| 169 |
"246": "NON_PARTICIPATING",
|
| 170 |
"247": "NON_PARTICIPATING",
|
|
|
|
| 175 |
"251": "NON_PARTICIPATING",
|
| 176 |
"252": "NON_PARTICIPATING",
|
| 177 |
"253": "NON_PARTICIPATING",
|
| 178 |
+
"254": "SUCCESS",
|
| 179 |
"255": "NON_PARTICIPATING",
|
| 180 |
"26": "SUCCESS",
|
| 181 |
"27": "SUCCESS",
|
| 182 |
+
"28": "SUCCESS",
|
| 183 |
"29": "SUCCESS",
|
| 184 |
"3": "SUCCESS",
|
| 185 |
"30": "NON_PARTICIPATING",
|
|
|
|
| 191 |
"36": "NON_PARTICIPATING",
|
| 192 |
"37": "SUCCESS",
|
| 193 |
"38": "NON_PARTICIPATING",
|
| 194 |
+
"39": "SUCCESS",
|
| 195 |
"4": "NON_PARTICIPATING",
|
| 196 |
"40": "SUCCESS",
|
| 197 |
"41": "SUCCESS",
|
|
|
|
| 223 |
"65": "SUCCESS",
|
| 224 |
"66": "NON_PARTICIPATING",
|
| 225 |
"67": "SUCCESS",
|
| 226 |
+
"68": "SUCCESS",
|
| 227 |
"69": "NON_PARTICIPATING",
|
| 228 |
"7": "SUCCESS",
|
| 229 |
"70": "NON_PARTICIPATING",
|
|
|
|
| 240 |
"80": "SUCCESS",
|
| 241 |
"81": "NON_PARTICIPATING",
|
| 242 |
"82": "SUCCESS",
|
| 243 |
+
"83": "SUCCESS",
|
| 244 |
"84": "SUCCESS",
|
| 245 |
"85": "NON_PARTICIPATING",
|
| 246 |
"86": "NON_PARTICIPATING",
|
|
|
|
| 250 |
"9": "NON_PARTICIPATING",
|
| 251 |
"90": "SUCCESS",
|
| 252 |
"91": "NON_PARTICIPATING",
|
| 253 |
+
"92": "SUCCESS",
|
| 254 |
"93": "NON_PARTICIPATING",
|
| 255 |
"94": "SUCCESS",
|
| 256 |
"95": "SUCCESS",
|
|
|
|
| 268 |
"AutoModelForCausalLM": "distributed/optimized-gpt2-500m--modeling_gpt_optimized.GPTOptim"
|
| 269 |
},
|
| 270 |
"block_list": [
|
| 271 |
+
5533454,
|
| 272 |
+
5533456,
|
| 273 |
+
5533458,
|
| 274 |
+
5533460,
|
| 275 |
+
5533462
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 276 |
],
|
| 277 |
"block_size": 1024,
|
| 278 |
"bos_token_id": 50256,
|
| 279 |
"embd_pdrop": 0.1,
|
| 280 |
"eos_token_id": 50256,
|
| 281 |
"initializer_range": 0.02,
|
| 282 |
+
"inner_step": 0,
|
| 283 |
"inner_steps": 0,
|
| 284 |
+
"last_allreduce_block": 5533305,
|
| 285 |
"layer_norm_epsilon": 1e-05,
|
| 286 |
"model_type": "gpt_optimized",
|
| 287 |
"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 8081782026
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6cd237108d9e69dca5213aeb40c54cf17be379f0a0b09d340bc484ae0dd07c2a
|
| 3 |
size 8081782026
|
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:a9cc9e913f798f6a2bb61366201917f03c4d0cb7daddd9d59b17f6736ca13bed
|
| 3 |
size 4040701744
|