Run 4. Outer Step 101. Inner Step 2.
Browse files- config.json +28 -28
- inner_optimizer.pt +1 -1
- model.safetensors +1 -1
config.json
CHANGED
|
@@ -6,10 +6,10 @@
|
|
| 6 |
"1": "SUCCESS",
|
| 7 |
"10": "NON_PARTICIPATING",
|
| 8 |
"100": "NON_PARTICIPATING",
|
| 9 |
-
"101": "
|
| 10 |
"102": "NON_PARTICIPATING",
|
| 11 |
"103": "NON_PARTICIPATING",
|
| 12 |
-
"104": "
|
| 13 |
"105": "NON_PARTICIPATING",
|
| 14 |
"106": "NON_PARTICIPATING",
|
| 15 |
"107": "NON_PARTICIPATING",
|
|
@@ -21,7 +21,7 @@
|
|
| 21 |
"112": "NON_PARTICIPATING",
|
| 22 |
"113": "NON_PARTICIPATING",
|
| 23 |
"114": "NON_PARTICIPATING",
|
| 24 |
-
"115": "
|
| 25 |
"116": "NON_PARTICIPATING",
|
| 26 |
"117": "NON_PARTICIPATING",
|
| 27 |
"118": "NON_PARTICIPATING",
|
|
@@ -46,13 +46,13 @@
|
|
| 46 |
"135": "NON_PARTICIPATING",
|
| 47 |
"136": "NON_PARTICIPATING",
|
| 48 |
"137": "NON_PARTICIPATING",
|
| 49 |
-
"138": "
|
| 50 |
"139": "NON_PARTICIPATING",
|
| 51 |
"14": "SUCCESS",
|
| 52 |
-
"140": "
|
| 53 |
"141": "NON_PARTICIPATING",
|
| 54 |
-
"142": "
|
| 55 |
-
"143": "
|
| 56 |
"144": "NON_PARTICIPATING",
|
| 57 |
"145": "NON_PARTICIPATING",
|
| 58 |
"146": "NON_PARTICIPATING",
|
|
@@ -66,7 +66,7 @@
|
|
| 66 |
"153": "NON_PARTICIPATING",
|
| 67 |
"154": "NON_PARTICIPATING",
|
| 68 |
"155": "NON_PARTICIPATING",
|
| 69 |
-
"156": "
|
| 70 |
"157": "NON_PARTICIPATING",
|
| 71 |
"158": "NON_PARTICIPATING",
|
| 72 |
"159": "NON_PARTICIPATING",
|
|
@@ -78,14 +78,14 @@
|
|
| 78 |
"164": "NON_PARTICIPATING",
|
| 79 |
"165": "NON_PARTICIPATING",
|
| 80 |
"166": "NON_PARTICIPATING",
|
| 81 |
-
"167": "
|
| 82 |
"168": "NON_PARTICIPATING",
|
| 83 |
"169": "NON_PARTICIPATING",
|
| 84 |
"17": "SUCCESS",
|
| 85 |
"170": "NON_PARTICIPATING",
|
| 86 |
"171": "NON_PARTICIPATING",
|
| 87 |
"172": "NON_PARTICIPATING",
|
| 88 |
-
"173": "
|
| 89 |
"174": "NON_PARTICIPATING",
|
| 90 |
"175": "NON_PARTICIPATING",
|
| 91 |
"176": "NON_PARTICIPATING",
|
|
@@ -93,7 +93,7 @@
|
|
| 93 |
"178": "NON_PARTICIPATING",
|
| 94 |
"179": "SUCCESS",
|
| 95 |
"18": "SUCCESS",
|
| 96 |
-
"180": "
|
| 97 |
"181": "NON_PARTICIPATING",
|
| 98 |
"182": "NON_PARTICIPATING",
|
| 99 |
"183": "NON_PARTICIPATING",
|
|
@@ -108,20 +108,20 @@
|
|
| 108 |
"191": "NON_PARTICIPATING",
|
| 109 |
"192": "NON_PARTICIPATING",
|
| 110 |
"193": "NON_PARTICIPATING",
|
| 111 |
-
"194": "
|
| 112 |
"195": "NON_PARTICIPATING",
|
| 113 |
"196": "NON_PARTICIPATING",
|
| 114 |
"197": "NON_PARTICIPATING",
|
| 115 |
"198": "NON_PARTICIPATING",
|
| 116 |
"199": "NON_PARTICIPATING",
|
| 117 |
"2": "SUCCESS",
|
| 118 |
-
"20": "
|
| 119 |
-
"200": "
|
| 120 |
"201": "NON_PARTICIPATING",
|
| 121 |
"202": "NON_PARTICIPATING",
|
| 122 |
-
"203": "
|
| 123 |
"204": "NON_PARTICIPATING",
|
| 124 |
-
"205": "
|
| 125 |
"206": "NON_PARTICIPATING",
|
| 126 |
"207": "NON_PARTICIPATING",
|
| 127 |
"208": "NON_PARTICIPATING",
|
|
@@ -151,11 +151,11 @@
|
|
| 151 |
"23": "SUCCESS",
|
| 152 |
"230": "NON_PARTICIPATING",
|
| 153 |
"231": "NON_PARTICIPATING",
|
| 154 |
-
"232": "
|
| 155 |
"233": "NON_PARTICIPATING",
|
| 156 |
"234": "NON_PARTICIPATING",
|
| 157 |
"235": "NON_PARTICIPATING",
|
| 158 |
-
"236": "
|
| 159 |
"237": "NON_PARTICIPATING",
|
| 160 |
"238": "NON_PARTICIPATING",
|
| 161 |
"239": "NON_PARTICIPATING",
|
|
@@ -168,7 +168,7 @@
|
|
| 168 |
"245": "NON_PARTICIPATING",
|
| 169 |
"246": "NON_PARTICIPATING",
|
| 170 |
"247": "NON_PARTICIPATING",
|
| 171 |
-
"248": "
|
| 172 |
"249": "NON_PARTICIPATING",
|
| 173 |
"25": "SUCCESS",
|
| 174 |
"250": "NON_PARTICIPATING",
|
|
@@ -197,12 +197,12 @@
|
|
| 197 |
"41": "NON_PARTICIPATING",
|
| 198 |
"42": "NON_PARTICIPATING",
|
| 199 |
"43": "NON_PARTICIPATING",
|
| 200 |
-
"44": "
|
| 201 |
"45": "SUCCESS",
|
| 202 |
"46": "SUCCESS",
|
| 203 |
"47": "NON_PARTICIPATING",
|
| 204 |
"48": "NON_PARTICIPATING",
|
| 205 |
-
"49": "
|
| 206 |
"5": "NON_PARTICIPATING",
|
| 207 |
"50": "SUCCESS",
|
| 208 |
"51": "SUCCESS",
|
|
@@ -227,13 +227,13 @@
|
|
| 227 |
"69": "SUCCESS",
|
| 228 |
"7": "SUCCESS",
|
| 229 |
"70": "NON_PARTICIPATING",
|
| 230 |
-
"71": "
|
| 231 |
"72": "NON_PARTICIPATING",
|
| 232 |
"73": "SUCCESS",
|
| 233 |
"74": "NON_PARTICIPATING",
|
| 234 |
"75": "NON_PARTICIPATING",
|
| 235 |
"76": "SUCCESS",
|
| 236 |
-
"77": "
|
| 237 |
"78": "SUCCESS",
|
| 238 |
"79": "NON_PARTICIPATING",
|
| 239 |
"8": "NON_PARTICIPATING",
|
|
@@ -253,7 +253,7 @@
|
|
| 253 |
"92": "NON_PARTICIPATING",
|
| 254 |
"93": "SUCCESS",
|
| 255 |
"94": "SUCCESS",
|
| 256 |
-
"95": "
|
| 257 |
"96": "NON_PARTICIPATING",
|
| 258 |
"97": "NON_PARTICIPATING",
|
| 259 |
"98": "SUCCESS",
|
|
@@ -268,17 +268,17 @@
|
|
| 268 |
"AutoModelForCausalLM": "distributed/optimized-gpt2-500m--modeling_gpt_optimized.GPTOptim"
|
| 269 |
},
|
| 270 |
"block_list": [
|
| 271 |
-
|
| 272 |
-
|
| 273 |
],
|
| 274 |
"block_size": 1024,
|
| 275 |
"bos_token_id": 50256,
|
| 276 |
"embd_pdrop": 0.1,
|
| 277 |
"eos_token_id": 50256,
|
| 278 |
"initializer_range": 0.02,
|
| 279 |
-
"inner_step":
|
| 280 |
"inner_steps": 0,
|
| 281 |
-
"last_allreduce_block":
|
| 282 |
"layer_norm_epsilon": 1e-05,
|
| 283 |
"model_type": "gpt_optimized",
|
| 284 |
"n_embd": 1280,
|
|
|
|
| 6 |
"1": "SUCCESS",
|
| 7 |
"10": "NON_PARTICIPATING",
|
| 8 |
"100": "NON_PARTICIPATING",
|
| 9 |
+
"101": "NON_PARTICIPATING",
|
| 10 |
"102": "NON_PARTICIPATING",
|
| 11 |
"103": "NON_PARTICIPATING",
|
| 12 |
+
"104": "SUCCESS",
|
| 13 |
"105": "NON_PARTICIPATING",
|
| 14 |
"106": "NON_PARTICIPATING",
|
| 15 |
"107": "NON_PARTICIPATING",
|
|
|
|
| 21 |
"112": "NON_PARTICIPATING",
|
| 22 |
"113": "NON_PARTICIPATING",
|
| 23 |
"114": "NON_PARTICIPATING",
|
| 24 |
+
"115": "SUCCESS",
|
| 25 |
"116": "NON_PARTICIPATING",
|
| 26 |
"117": "NON_PARTICIPATING",
|
| 27 |
"118": "NON_PARTICIPATING",
|
|
|
|
| 46 |
"135": "NON_PARTICIPATING",
|
| 47 |
"136": "NON_PARTICIPATING",
|
| 48 |
"137": "NON_PARTICIPATING",
|
| 49 |
+
"138": "NON_PARTICIPATING",
|
| 50 |
"139": "NON_PARTICIPATING",
|
| 51 |
"14": "SUCCESS",
|
| 52 |
+
"140": "SUCCESS",
|
| 53 |
"141": "NON_PARTICIPATING",
|
| 54 |
+
"142": "FAIL",
|
| 55 |
+
"143": "SUCCESS",
|
| 56 |
"144": "NON_PARTICIPATING",
|
| 57 |
"145": "NON_PARTICIPATING",
|
| 58 |
"146": "NON_PARTICIPATING",
|
|
|
|
| 66 |
"153": "NON_PARTICIPATING",
|
| 67 |
"154": "NON_PARTICIPATING",
|
| 68 |
"155": "NON_PARTICIPATING",
|
| 69 |
+
"156": "SUCCESS",
|
| 70 |
"157": "NON_PARTICIPATING",
|
| 71 |
"158": "NON_PARTICIPATING",
|
| 72 |
"159": "NON_PARTICIPATING",
|
|
|
|
| 78 |
"164": "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": "NON_PARTICIPATING",
|
| 86 |
"171": "NON_PARTICIPATING",
|
| 87 |
"172": "NON_PARTICIPATING",
|
| 88 |
+
"173": "NON_PARTICIPATING",
|
| 89 |
"174": "NON_PARTICIPATING",
|
| 90 |
"175": "NON_PARTICIPATING",
|
| 91 |
"176": "NON_PARTICIPATING",
|
|
|
|
| 93 |
"178": "NON_PARTICIPATING",
|
| 94 |
"179": "SUCCESS",
|
| 95 |
"18": "SUCCESS",
|
| 96 |
+
"180": "SUCCESS",
|
| 97 |
"181": "NON_PARTICIPATING",
|
| 98 |
"182": "NON_PARTICIPATING",
|
| 99 |
"183": "NON_PARTICIPATING",
|
|
|
|
| 108 |
"191": "NON_PARTICIPATING",
|
| 109 |
"192": "NON_PARTICIPATING",
|
| 110 |
"193": "NON_PARTICIPATING",
|
| 111 |
+
"194": "SUCCESS",
|
| 112 |
"195": "NON_PARTICIPATING",
|
| 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",
|
| 123 |
"204": "NON_PARTICIPATING",
|
| 124 |
+
"205": "NON_PARTICIPATING",
|
| 125 |
"206": "NON_PARTICIPATING",
|
| 126 |
"207": "NON_PARTICIPATING",
|
| 127 |
"208": "NON_PARTICIPATING",
|
|
|
|
| 151 |
"23": "SUCCESS",
|
| 152 |
"230": "NON_PARTICIPATING",
|
| 153 |
"231": "NON_PARTICIPATING",
|
| 154 |
+
"232": "NON_PARTICIPATING",
|
| 155 |
"233": "NON_PARTICIPATING",
|
| 156 |
"234": "NON_PARTICIPATING",
|
| 157 |
"235": "NON_PARTICIPATING",
|
| 158 |
+
"236": "NON_PARTICIPATING",
|
| 159 |
"237": "NON_PARTICIPATING",
|
| 160 |
"238": "NON_PARTICIPATING",
|
| 161 |
"239": "NON_PARTICIPATING",
|
|
|
|
| 168 |
"245": "NON_PARTICIPATING",
|
| 169 |
"246": "NON_PARTICIPATING",
|
| 170 |
"247": "NON_PARTICIPATING",
|
| 171 |
+
"248": "NON_PARTICIPATING",
|
| 172 |
"249": "NON_PARTICIPATING",
|
| 173 |
"25": "SUCCESS",
|
| 174 |
"250": "NON_PARTICIPATING",
|
|
|
|
| 197 |
"41": "NON_PARTICIPATING",
|
| 198 |
"42": "NON_PARTICIPATING",
|
| 199 |
"43": "NON_PARTICIPATING",
|
| 200 |
+
"44": "SUCCESS",
|
| 201 |
"45": "SUCCESS",
|
| 202 |
"46": "SUCCESS",
|
| 203 |
"47": "NON_PARTICIPATING",
|
| 204 |
"48": "NON_PARTICIPATING",
|
| 205 |
+
"49": "SUCCESS",
|
| 206 |
"5": "NON_PARTICIPATING",
|
| 207 |
"50": "SUCCESS",
|
| 208 |
"51": "SUCCESS",
|
|
|
|
| 227 |
"69": "SUCCESS",
|
| 228 |
"7": "SUCCESS",
|
| 229 |
"70": "NON_PARTICIPATING",
|
| 230 |
+
"71": "SUCCESS",
|
| 231 |
"72": "NON_PARTICIPATING",
|
| 232 |
"73": "SUCCESS",
|
| 233 |
"74": "NON_PARTICIPATING",
|
| 234 |
"75": "NON_PARTICIPATING",
|
| 235 |
"76": "SUCCESS",
|
| 236 |
+
"77": "SUCCESS",
|
| 237 |
"78": "SUCCESS",
|
| 238 |
"79": "NON_PARTICIPATING",
|
| 239 |
"8": "NON_PARTICIPATING",
|
|
|
|
| 253 |
"92": "NON_PARTICIPATING",
|
| 254 |
"93": "SUCCESS",
|
| 255 |
"94": "SUCCESS",
|
| 256 |
+
"95": "SUCCESS",
|
| 257 |
"96": "NON_PARTICIPATING",
|
| 258 |
"97": "NON_PARTICIPATING",
|
| 259 |
"98": "SUCCESS",
|
|
|
|
| 268 |
"AutoModelForCausalLM": "distributed/optimized-gpt2-500m--modeling_gpt_optimized.GPTOptim"
|
| 269 |
},
|
| 270 |
"block_list": [
|
| 271 |
+
5697011,
|
| 272 |
+
5697041
|
| 273 |
],
|
| 274 |
"block_size": 1024,
|
| 275 |
"bos_token_id": 50256,
|
| 276 |
"embd_pdrop": 0.1,
|
| 277 |
"eos_token_id": 50256,
|
| 278 |
"initializer_range": 0.02,
|
| 279 |
+
"inner_step": 2,
|
| 280 |
"inner_steps": 0,
|
| 281 |
+
"last_allreduce_block": 5696880,
|
| 282 |
"layer_norm_epsilon": 1e-05,
|
| 283 |
"model_type": "gpt_optimized",
|
| 284 |
"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:b32f6cfc65ac6d1ae379abf7fe8a526b666f764670087230605b09a20e18f563
|
| 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 sha256:5a6cf121bd29ce006e68b464603c6b094241090f0a31aeb695baf00019ad52e5
|
| 3 |
size 4040701744
|