Run 4. Outer Step 55. Inner Step 2.
Browse files- config.json +43 -46
- inner_optimizer.pt +1 -1
- model.safetensors +1 -1
config.json
CHANGED
|
@@ -1,72 +1,72 @@
|
|
| 1 |
{
|
| 2 |
-
"_name_or_path": "
|
| 3 |
"activation_function": "gelu_new",
|
| 4 |
"all_reduce_scores": {
|
| 5 |
"0": "NON_PARTICIPATING",
|
| 6 |
-
"1": "
|
| 7 |
"10": "NON_PARTICIPATING",
|
| 8 |
"100": "NON_PARTICIPATING",
|
| 9 |
-
"101": "
|
| 10 |
"102": "NON_PARTICIPATING",
|
| 11 |
"103": "NON_PARTICIPATING",
|
| 12 |
"104": "NON_PARTICIPATING",
|
| 13 |
"105": "NON_PARTICIPATING",
|
| 14 |
-
"106": "
|
| 15 |
"107": "SUCCESS",
|
| 16 |
-
"108": "
|
| 17 |
"109": "NON_PARTICIPATING",
|
| 18 |
"11": "SUCCESS",
|
| 19 |
-
"110": "
|
| 20 |
"111": "NON_PARTICIPATING",
|
| 21 |
"112": "NON_PARTICIPATING",
|
| 22 |
"113": "NON_PARTICIPATING",
|
| 23 |
"114": "NON_PARTICIPATING",
|
| 24 |
"115": "NON_PARTICIPATING",
|
| 25 |
"116": "NON_PARTICIPATING",
|
| 26 |
-
"117": "
|
| 27 |
"118": "NON_PARTICIPATING",
|
| 28 |
"119": "NON_PARTICIPATING",
|
| 29 |
"12": "SUCCESS",
|
| 30 |
-
"120": "
|
| 31 |
"121": "NON_PARTICIPATING",
|
| 32 |
"122": "SUCCESS",
|
| 33 |
"123": "NON_PARTICIPATING",
|
| 34 |
"124": "NON_PARTICIPATING",
|
| 35 |
-
"125": "
|
| 36 |
-
"126": "
|
| 37 |
"127": "NON_PARTICIPATING",
|
| 38 |
"128": "NON_PARTICIPATING",
|
| 39 |
"129": "NON_PARTICIPATING",
|
| 40 |
"13": "SUCCESS",
|
| 41 |
-
"130": "
|
| 42 |
"131": "NON_PARTICIPATING",
|
| 43 |
"132": "NON_PARTICIPATING",
|
| 44 |
"133": "NON_PARTICIPATING",
|
| 45 |
-
"134": "
|
| 46 |
"135": "NON_PARTICIPATING",
|
| 47 |
"136": "NON_PARTICIPATING",
|
| 48 |
"137": "NON_PARTICIPATING",
|
| 49 |
"138": "NON_PARTICIPATING",
|
| 50 |
-
"139": "
|
| 51 |
"14": "NON_PARTICIPATING",
|
| 52 |
"140": "NON_PARTICIPATING",
|
| 53 |
-
"141": "
|
| 54 |
"142": "NON_PARTICIPATING",
|
| 55 |
-
"143": "
|
| 56 |
"144": "NON_PARTICIPATING",
|
| 57 |
"145": "NON_PARTICIPATING",
|
| 58 |
"146": "NON_PARTICIPATING",
|
| 59 |
"147": "NON_PARTICIPATING",
|
| 60 |
"148": "NON_PARTICIPATING",
|
| 61 |
"149": "NON_PARTICIPATING",
|
| 62 |
-
"15": "
|
| 63 |
-
"150": "
|
| 64 |
"151": "NON_PARTICIPATING",
|
| 65 |
"152": "NON_PARTICIPATING",
|
| 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",
|
|
@@ -83,24 +83,24 @@
|
|
| 83 |
"169": "NON_PARTICIPATING",
|
| 84 |
"17": "SUCCESS",
|
| 85 |
"170": "NON_PARTICIPATING",
|
| 86 |
-
"171": "
|
| 87 |
"172": "NON_PARTICIPATING",
|
| 88 |
-
"173": "
|
| 89 |
"174": "NON_PARTICIPATING",
|
| 90 |
"175": "NON_PARTICIPATING",
|
| 91 |
"176": "NON_PARTICIPATING",
|
| 92 |
-
"177": "
|
| 93 |
-
"178": "
|
| 94 |
"179": "NON_PARTICIPATING",
|
| 95 |
"18": "SUCCESS",
|
| 96 |
"180": "NON_PARTICIPATING",
|
| 97 |
"181": "SUCCESS",
|
| 98 |
-
"182": "
|
| 99 |
"183": "NON_PARTICIPATING",
|
| 100 |
-
"184": "
|
| 101 |
"185": "NON_PARTICIPATING",
|
| 102 |
"186": "NON_PARTICIPATING",
|
| 103 |
-
"187": "
|
| 104 |
"188": "NON_PARTICIPATING",
|
| 105 |
"189": "NON_PARTICIPATING",
|
| 106 |
"19": "NON_PARTICIPATING",
|
|
@@ -124,7 +124,7 @@
|
|
| 124 |
"205": "NON_PARTICIPATING",
|
| 125 |
"206": "NON_PARTICIPATING",
|
| 126 |
"207": "NON_PARTICIPATING",
|
| 127 |
-
"208": "
|
| 128 |
"209": "NON_PARTICIPATING",
|
| 129 |
"21": "SUCCESS",
|
| 130 |
"210": "NON_PARTICIPATING",
|
|
@@ -133,14 +133,14 @@
|
|
| 133 |
"213": "NON_PARTICIPATING",
|
| 134 |
"214": "NON_PARTICIPATING",
|
| 135 |
"215": "NON_PARTICIPATING",
|
| 136 |
-
"216": "
|
| 137 |
-
"217": "
|
| 138 |
-
"218": "
|
| 139 |
-
"219": "
|
| 140 |
"22": "SUCCESS",
|
| 141 |
"220": "NON_PARTICIPATING",
|
| 142 |
"221": "NON_PARTICIPATING",
|
| 143 |
-
"222": "
|
| 144 |
"223": "NON_PARTICIPATING",
|
| 145 |
"224": "NON_PARTICIPATING",
|
| 146 |
"225": "NON_PARTICIPATING",
|
|
@@ -149,13 +149,13 @@
|
|
| 149 |
"228": "NON_PARTICIPATING",
|
| 150 |
"229": "NON_PARTICIPATING",
|
| 151 |
"23": "SUCCESS",
|
| 152 |
-
"230": "
|
| 153 |
"231": "NON_PARTICIPATING",
|
| 154 |
"232": "NON_PARTICIPATING",
|
| 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",
|
|
@@ -175,7 +175,7 @@
|
|
| 175 |
"251": "NON_PARTICIPATING",
|
| 176 |
"252": "NON_PARTICIPATING",
|
| 177 |
"253": "NON_PARTICIPATING",
|
| 178 |
-
"254": "
|
| 179 |
"255": "NON_PARTICIPATING",
|
| 180 |
"26": "SUCCESS",
|
| 181 |
"27": "SUCCESS",
|
|
@@ -196,7 +196,7 @@
|
|
| 196 |
"40": "NON_PARTICIPATING",
|
| 197 |
"41": "SUCCESS",
|
| 198 |
"42": "SUCCESS",
|
| 199 |
-
"43": "
|
| 200 |
"44": "SUCCESS",
|
| 201 |
"45": "SUCCESS",
|
| 202 |
"46": "NON_PARTICIPATING",
|
|
@@ -215,7 +215,7 @@
|
|
| 215 |
"58": "NON_PARTICIPATING",
|
| 216 |
"59": "NON_PARTICIPATING",
|
| 217 |
"6": "NON_PARTICIPATING",
|
| 218 |
-
"60": "
|
| 219 |
"61": "NON_PARTICIPATING",
|
| 220 |
"62": "SUCCESS",
|
| 221 |
"63": "SUCCESS",
|
|
@@ -227,20 +227,20 @@
|
|
| 227 |
"69": "SUCCESS",
|
| 228 |
"7": "SUCCESS",
|
| 229 |
"70": "SUCCESS",
|
| 230 |
-
"71": "
|
| 231 |
"72": "SUCCESS",
|
| 232 |
"73": "NON_PARTICIPATING",
|
| 233 |
"74": "SUCCESS",
|
| 234 |
"75": "SUCCESS",
|
| 235 |
"76": "NON_PARTICIPATING",
|
| 236 |
-
"77": "
|
| 237 |
"78": "SUCCESS",
|
| 238 |
"79": "NON_PARTICIPATING",
|
| 239 |
"8": "NON_PARTICIPATING",
|
| 240 |
"80": "SUCCESS",
|
| 241 |
"81": "SUCCESS",
|
| 242 |
"82": "SUCCESS",
|
| 243 |
-
"83": "
|
| 244 |
"84": "SUCCESS",
|
| 245 |
"85": "SUCCESS",
|
| 246 |
"86": "NON_PARTICIPATING",
|
|
@@ -250,7 +250,7 @@
|
|
| 250 |
"9": "SUCCESS",
|
| 251 |
"90": "SUCCESS",
|
| 252 |
"91": "NON_PARTICIPATING",
|
| 253 |
-
"92": "
|
| 254 |
"93": "NON_PARTICIPATING",
|
| 255 |
"94": "SUCCESS",
|
| 256 |
"95": "SUCCESS",
|
|
@@ -267,18 +267,15 @@
|
|
| 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 |
-
5632746,
|
| 272 |
-
5632775
|
| 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,
|
|
|
|
| 1 |
{
|
| 2 |
+
"_name_or_path": "distributed/optimized-gpt2-1b",
|
| 3 |
"activation_function": "gelu_new",
|
| 4 |
"all_reduce_scores": {
|
| 5 |
"0": "NON_PARTICIPATING",
|
| 6 |
+
"1": "SUCCESS",
|
| 7 |
"10": "NON_PARTICIPATING",
|
| 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": "SUCCESS",
|
| 16 |
+
"108": "NON_PARTICIPATING",
|
| 17 |
"109": "NON_PARTICIPATING",
|
| 18 |
"11": "SUCCESS",
|
| 19 |
+
"110": "NON_PARTICIPATING",
|
| 20 |
"111": "NON_PARTICIPATING",
|
| 21 |
"112": "NON_PARTICIPATING",
|
| 22 |
"113": "NON_PARTICIPATING",
|
| 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": "NON_PARTICIPATING",
|
| 32 |
"122": "SUCCESS",
|
| 33 |
"123": "NON_PARTICIPATING",
|
| 34 |
"124": "NON_PARTICIPATING",
|
| 35 |
+
"125": "NON_PARTICIPATING",
|
| 36 |
+
"126": "SUCCESS",
|
| 37 |
"127": "NON_PARTICIPATING",
|
| 38 |
"128": "NON_PARTICIPATING",
|
| 39 |
"129": "NON_PARTICIPATING",
|
| 40 |
"13": "SUCCESS",
|
| 41 |
+
"130": "SUCCESS",
|
| 42 |
"131": "NON_PARTICIPATING",
|
| 43 |
"132": "NON_PARTICIPATING",
|
| 44 |
"133": "NON_PARTICIPATING",
|
| 45 |
+
"134": "NON_PARTICIPATING",
|
| 46 |
"135": "NON_PARTICIPATING",
|
| 47 |
"136": "NON_PARTICIPATING",
|
| 48 |
"137": "NON_PARTICIPATING",
|
| 49 |
"138": "NON_PARTICIPATING",
|
| 50 |
+
"139": "SUCCESS",
|
| 51 |
"14": "NON_PARTICIPATING",
|
| 52 |
"140": "NON_PARTICIPATING",
|
| 53 |
+
"141": "NON_PARTICIPATING",
|
| 54 |
"142": "NON_PARTICIPATING",
|
| 55 |
+
"143": "SUCCESS",
|
| 56 |
"144": "NON_PARTICIPATING",
|
| 57 |
"145": "NON_PARTICIPATING",
|
| 58 |
"146": "NON_PARTICIPATING",
|
| 59 |
"147": "NON_PARTICIPATING",
|
| 60 |
"148": "NON_PARTICIPATING",
|
| 61 |
"149": "NON_PARTICIPATING",
|
| 62 |
+
"15": "SUCCESS",
|
| 63 |
+
"150": "NON_PARTICIPATING",
|
| 64 |
"151": "NON_PARTICIPATING",
|
| 65 |
"152": "NON_PARTICIPATING",
|
| 66 |
"153": "NON_PARTICIPATING",
|
| 67 |
"154": "NON_PARTICIPATING",
|
| 68 |
"155": "NON_PARTICIPATING",
|
| 69 |
+
"156": "NON_PARTICIPATING",
|
| 70 |
"157": "NON_PARTICIPATING",
|
| 71 |
"158": "NON_PARTICIPATING",
|
| 72 |
"159": "NON_PARTICIPATING",
|
|
|
|
| 83 |
"169": "NON_PARTICIPATING",
|
| 84 |
"17": "SUCCESS",
|
| 85 |
"170": "NON_PARTICIPATING",
|
| 86 |
+
"171": "SUCCESS",
|
| 87 |
"172": "NON_PARTICIPATING",
|
| 88 |
+
"173": "SUCCESS",
|
| 89 |
"174": "NON_PARTICIPATING",
|
| 90 |
"175": "NON_PARTICIPATING",
|
| 91 |
"176": "NON_PARTICIPATING",
|
| 92 |
+
"177": "SUCCESS",
|
| 93 |
+
"178": "NON_PARTICIPATING",
|
| 94 |
"179": "NON_PARTICIPATING",
|
| 95 |
"18": "SUCCESS",
|
| 96 |
"180": "NON_PARTICIPATING",
|
| 97 |
"181": "SUCCESS",
|
| 98 |
+
"182": "SUCCESS",
|
| 99 |
"183": "NON_PARTICIPATING",
|
| 100 |
+
"184": "SUCCESS",
|
| 101 |
"185": "NON_PARTICIPATING",
|
| 102 |
"186": "NON_PARTICIPATING",
|
| 103 |
+
"187": "NON_PARTICIPATING",
|
| 104 |
"188": "NON_PARTICIPATING",
|
| 105 |
"189": "NON_PARTICIPATING",
|
| 106 |
"19": "NON_PARTICIPATING",
|
|
|
|
| 124 |
"205": "NON_PARTICIPATING",
|
| 125 |
"206": "NON_PARTICIPATING",
|
| 126 |
"207": "NON_PARTICIPATING",
|
| 127 |
+
"208": "SUCCESS",
|
| 128 |
"209": "NON_PARTICIPATING",
|
| 129 |
"21": "SUCCESS",
|
| 130 |
"210": "NON_PARTICIPATING",
|
|
|
|
| 133 |
"213": "NON_PARTICIPATING",
|
| 134 |
"214": "NON_PARTICIPATING",
|
| 135 |
"215": "NON_PARTICIPATING",
|
| 136 |
+
"216": "NON_PARTICIPATING",
|
| 137 |
+
"217": "SUCCESS",
|
| 138 |
+
"218": "NON_PARTICIPATING",
|
| 139 |
+
"219": "NON_PARTICIPATING",
|
| 140 |
"22": "SUCCESS",
|
| 141 |
"220": "NON_PARTICIPATING",
|
| 142 |
"221": "NON_PARTICIPATING",
|
| 143 |
+
"222": "NON_PARTICIPATING",
|
| 144 |
"223": "NON_PARTICIPATING",
|
| 145 |
"224": "NON_PARTICIPATING",
|
| 146 |
"225": "NON_PARTICIPATING",
|
|
|
|
| 149 |
"228": "NON_PARTICIPATING",
|
| 150 |
"229": "NON_PARTICIPATING",
|
| 151 |
"23": "SUCCESS",
|
| 152 |
+
"230": "SUCCESS",
|
| 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",
|
|
|
|
| 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",
|
|
|
|
| 196 |
"40": "NON_PARTICIPATING",
|
| 197 |
"41": "SUCCESS",
|
| 198 |
"42": "SUCCESS",
|
| 199 |
+
"43": "NON_PARTICIPATING",
|
| 200 |
"44": "SUCCESS",
|
| 201 |
"45": "SUCCESS",
|
| 202 |
"46": "NON_PARTICIPATING",
|
|
|
|
| 215 |
"58": "NON_PARTICIPATING",
|
| 216 |
"59": "NON_PARTICIPATING",
|
| 217 |
"6": "NON_PARTICIPATING",
|
| 218 |
+
"60": "SUCCESS",
|
| 219 |
"61": "NON_PARTICIPATING",
|
| 220 |
"62": "SUCCESS",
|
| 221 |
"63": "SUCCESS",
|
|
|
|
| 227 |
"69": "SUCCESS",
|
| 228 |
"7": "SUCCESS",
|
| 229 |
"70": "SUCCESS",
|
| 230 |
+
"71": "SUCCESS",
|
| 231 |
"72": "SUCCESS",
|
| 232 |
"73": "NON_PARTICIPATING",
|
| 233 |
"74": "SUCCESS",
|
| 234 |
"75": "SUCCESS",
|
| 235 |
"76": "NON_PARTICIPATING",
|
| 236 |
+
"77": "SUCCESS",
|
| 237 |
"78": "SUCCESS",
|
| 238 |
"79": "NON_PARTICIPATING",
|
| 239 |
"8": "NON_PARTICIPATING",
|
| 240 |
"80": "SUCCESS",
|
| 241 |
"81": "SUCCESS",
|
| 242 |
"82": "SUCCESS",
|
| 243 |
+
"83": "SUCCESS",
|
| 244 |
"84": "SUCCESS",
|
| 245 |
"85": "SUCCESS",
|
| 246 |
"86": "NON_PARTICIPATING",
|
|
|
|
| 250 |
"9": "SUCCESS",
|
| 251 |
"90": "SUCCESS",
|
| 252 |
"91": "NON_PARTICIPATING",
|
| 253 |
+
"92": "NON_PARTICIPATING",
|
| 254 |
"93": "NON_PARTICIPATING",
|
| 255 |
"94": "SUCCESS",
|
| 256 |
"95": "SUCCESS",
|
|
|
|
| 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 |
"block_size": 1024,
|
| 272 |
"bos_token_id": 50256,
|
| 273 |
"embd_pdrop": 0.1,
|
| 274 |
"eos_token_id": 50256,
|
| 275 |
"initializer_range": 0.02,
|
| 276 |
+
"inner_step": 2,
|
| 277 |
"inner_steps": 0,
|
| 278 |
+
"last_allreduce_block": 5632906,
|
| 279 |
"layer_norm_epsilon": 1e-05,
|
| 280 |
"model_type": "gpt_optimized",
|
| 281 |
"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 8081782503
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:562c2c84378a86ec2d5978670e12f2c378503f52d22f670ccfecdf243bd86c06
|
| 3 |
size 8081782503
|
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 sha256:022bc6ddb098beb89e6511ce417c696bf2b088d4866e20ce3766b2fb07c25546
|
| 3 |
size 4040701744
|