Upload folder using huggingface_hub (#85)
Browse files- 6d110a0d3e28613facbeaa283990ea3d249f3e345766d3de140b9b1d9ccb427a (ae70fbea1dd0cff5c4f12cbae4e72482b297f5e6)
- config.json +52 -52
- model.safetensors +1 -1
- smash_config.json +1 -1
config.json
CHANGED
|
@@ -1,5 +1,5 @@
|
|
| 1 |
{
|
| 2 |
-
"_name_or_path": "/covalent/.cache/models/
|
| 3 |
"activation_function": "gelu_new",
|
| 4 |
"all_reduce_scores": {
|
| 5 |
"0": "NON_PARTICIPATING",
|
|
@@ -9,9 +9,9 @@
|
|
| 9 |
"101": "SUCCESS",
|
| 10 |
"102": "NON_PARTICIPATING",
|
| 11 |
"103": "NON_PARTICIPATING",
|
| 12 |
-
"104": "
|
| 13 |
"105": "NON_PARTICIPATING",
|
| 14 |
-
"106": "
|
| 15 |
"107": "NON_PARTICIPATING",
|
| 16 |
"108": "SUCCESS",
|
| 17 |
"109": "NON_PARTICIPATING",
|
|
@@ -20,43 +20,43 @@
|
|
| 20 |
"111": "NON_PARTICIPATING",
|
| 21 |
"112": "NON_PARTICIPATING",
|
| 22 |
"113": "NON_PARTICIPATING",
|
| 23 |
-
"114": "
|
| 24 |
"115": "SUCCESS",
|
| 25 |
"116": "NON_PARTICIPATING",
|
| 26 |
-
"117": "
|
| 27 |
"118": "NON_PARTICIPATING",
|
| 28 |
"119": "SUCCESS",
|
| 29 |
-
"12": "
|
| 30 |
"120": "NON_PARTICIPATING",
|
| 31 |
"121": "SUCCESS",
|
| 32 |
-
"122": "
|
| 33 |
-
"123": "
|
| 34 |
"124": "SUCCESS",
|
| 35 |
"125": "NON_PARTICIPATING",
|
| 36 |
"126": "NON_PARTICIPATING",
|
| 37 |
"127": "NON_PARTICIPATING",
|
| 38 |
"128": "NON_PARTICIPATING",
|
| 39 |
"129": "SUCCESS",
|
| 40 |
-
"13": "
|
| 41 |
"130": "NON_PARTICIPATING",
|
| 42 |
"131": "NON_PARTICIPATING",
|
| 43 |
"132": "NON_PARTICIPATING",
|
| 44 |
"133": "NON_PARTICIPATING",
|
| 45 |
-
"134": "
|
| 46 |
-
"135": "
|
| 47 |
"136": "NON_PARTICIPATING",
|
| 48 |
-
"137": "
|
| 49 |
"138": "SUCCESS",
|
| 50 |
"139": "NON_PARTICIPATING",
|
| 51 |
"14": "SUCCESS",
|
| 52 |
"140": "NON_PARTICIPATING",
|
| 53 |
"141": "SUCCESS",
|
| 54 |
"142": "SUCCESS",
|
| 55 |
-
"143": "
|
| 56 |
"144": "NON_PARTICIPATING",
|
| 57 |
"145": "SUCCESS",
|
| 58 |
-
"146": "
|
| 59 |
-
"147": "
|
| 60 |
"148": "NON_PARTICIPATING",
|
| 61 |
"149": "NON_PARTICIPATING",
|
| 62 |
"15": "NON_PARTICIPATING",
|
|
@@ -64,13 +64,13 @@
|
|
| 64 |
"151": "SUCCESS",
|
| 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": "SUCCESS",
|
| 73 |
-
"16": "
|
| 74 |
"160": "NON_PARTICIPATING",
|
| 75 |
"161": "NON_PARTICIPATING",
|
| 76 |
"162": "NON_PARTICIPATING",
|
|
@@ -78,8 +78,8 @@
|
|
| 78 |
"164": "SUCCESS",
|
| 79 |
"165": "NON_PARTICIPATING",
|
| 80 |
"166": "NON_PARTICIPATING",
|
| 81 |
-
"167": "
|
| 82 |
-
"168": "
|
| 83 |
"169": "NON_PARTICIPATING",
|
| 84 |
"17": "NON_PARTICIPATING",
|
| 85 |
"170": "NON_PARTICIPATING",
|
|
@@ -94,7 +94,7 @@
|
|
| 94 |
"179": "SUCCESS",
|
| 95 |
"18": "NON_PARTICIPATING",
|
| 96 |
"180": "NON_PARTICIPATING",
|
| 97 |
-
"181": "
|
| 98 |
"182": "NON_PARTICIPATING",
|
| 99 |
"183": "NON_PARTICIPATING",
|
| 100 |
"184": "NON_PARTICIPATING",
|
|
@@ -114,27 +114,27 @@
|
|
| 114 |
"197": "NON_PARTICIPATING",
|
| 115 |
"198": "NON_PARTICIPATING",
|
| 116 |
"199": "NON_PARTICIPATING",
|
| 117 |
-
"2": "
|
| 118 |
"20": "NON_PARTICIPATING",
|
| 119 |
"200": "NON_PARTICIPATING",
|
| 120 |
"201": "SUCCESS",
|
| 121 |
"202": "SUCCESS",
|
| 122 |
"203": "NON_PARTICIPATING",
|
| 123 |
-
"204": "
|
| 124 |
"205": "SUCCESS",
|
| 125 |
"206": "NON_PARTICIPATING",
|
| 126 |
"207": "NON_PARTICIPATING",
|
| 127 |
-
"208": "
|
| 128 |
-
"209": "
|
| 129 |
-
"21": "
|
| 130 |
"210": "NON_PARTICIPATING",
|
| 131 |
"211": "NON_PARTICIPATING",
|
| 132 |
"212": "NON_PARTICIPATING",
|
| 133 |
-
"213": "
|
| 134 |
"214": "NON_PARTICIPATING",
|
| 135 |
"215": "NON_PARTICIPATING",
|
| 136 |
-
"216": "
|
| 137 |
-
"217": "
|
| 138 |
"218": "NON_PARTICIPATING",
|
| 139 |
"219": "NON_PARTICIPATING",
|
| 140 |
"22": "SUCCESS",
|
|
@@ -143,7 +143,7 @@
|
|
| 143 |
"222": "NON_PARTICIPATING",
|
| 144 |
"223": "NON_PARTICIPATING",
|
| 145 |
"224": "NON_PARTICIPATING",
|
| 146 |
-
"225": "
|
| 147 |
"226": "NON_PARTICIPATING",
|
| 148 |
"227": "NON_PARTICIPATING",
|
| 149 |
"228": "NON_PARTICIPATING",
|
|
@@ -151,20 +151,20 @@
|
|
| 151 |
"23": "SUCCESS",
|
| 152 |
"230": "NON_PARTICIPATING",
|
| 153 |
"231": "NON_PARTICIPATING",
|
| 154 |
-
"232": "
|
| 155 |
"233": "NON_PARTICIPATING",
|
| 156 |
"234": "NON_PARTICIPATING",
|
| 157 |
-
"235": "
|
| 158 |
-
"236": "
|
| 159 |
"237": "NON_PARTICIPATING",
|
| 160 |
"238": "NON_PARTICIPATING",
|
| 161 |
-
"239": "
|
| 162 |
"24": "NON_PARTICIPATING",
|
| 163 |
"240": "NON_PARTICIPATING",
|
| 164 |
"241": "NON_PARTICIPATING",
|
| 165 |
"242": "NON_PARTICIPATING",
|
| 166 |
"243": "NON_PARTICIPATING",
|
| 167 |
-
"244": "
|
| 168 |
"245": "NON_PARTICIPATING",
|
| 169 |
"246": "NON_PARTICIPATING",
|
| 170 |
"247": "NON_PARTICIPATING",
|
|
@@ -178,7 +178,7 @@
|
|
| 178 |
"254": "NON_PARTICIPATING",
|
| 179 |
"255": "NON_PARTICIPATING",
|
| 180 |
"26": "NON_PARTICIPATING",
|
| 181 |
-
"27": "
|
| 182 |
"28": "SUCCESS",
|
| 183 |
"29": "NON_PARTICIPATING",
|
| 184 |
"3": "NON_PARTICIPATING",
|
|
@@ -206,23 +206,23 @@
|
|
| 206 |
"5": "NON_PARTICIPATING",
|
| 207 |
"50": "SUCCESS",
|
| 208 |
"51": "NON_PARTICIPATING",
|
| 209 |
-
"52": "
|
| 210 |
"53": "NON_PARTICIPATING",
|
| 211 |
"54": "SUCCESS",
|
| 212 |
"55": "SUCCESS",
|
| 213 |
-
"56": "
|
| 214 |
"57": "NON_PARTICIPATING",
|
| 215 |
"58": "NON_PARTICIPATING",
|
| 216 |
"59": "NON_PARTICIPATING",
|
| 217 |
-
"6": "
|
| 218 |
"60": "NON_PARTICIPATING",
|
| 219 |
"61": "NON_PARTICIPATING",
|
| 220 |
-
"62": "
|
| 221 |
-
"63": "
|
| 222 |
"64": "NON_PARTICIPATING",
|
| 223 |
"65": "NON_PARTICIPATING",
|
| 224 |
-
"66": "
|
| 225 |
-
"67": "
|
| 226 |
"68": "NON_PARTICIPATING",
|
| 227 |
"69": "NON_PARTICIPATING",
|
| 228 |
"7": "NON_PARTICIPATING",
|
|
@@ -233,30 +233,30 @@
|
|
| 233 |
"74": "SUCCESS",
|
| 234 |
"75": "SUCCESS",
|
| 235 |
"76": "NON_PARTICIPATING",
|
| 236 |
-
"77": "
|
| 237 |
"78": "SUCCESS",
|
| 238 |
-
"79": "
|
| 239 |
"8": "NON_PARTICIPATING",
|
| 240 |
"80": "NON_PARTICIPATING",
|
| 241 |
"81": "NON_PARTICIPATING",
|
| 242 |
-
"82": "
|
| 243 |
"83": "NON_PARTICIPATING",
|
| 244 |
-
"84": "
|
| 245 |
"85": "NON_PARTICIPATING",
|
| 246 |
-
"86": "
|
| 247 |
"87": "NON_PARTICIPATING",
|
| 248 |
"88": "SUCCESS",
|
| 249 |
"89": "NON_PARTICIPATING",
|
| 250 |
-
"9": "
|
| 251 |
"90": "NON_PARTICIPATING",
|
| 252 |
"91": "NON_PARTICIPATING",
|
| 253 |
"92": "NON_PARTICIPATING",
|
| 254 |
-
"93": "
|
| 255 |
"94": "NON_PARTICIPATING",
|
| 256 |
-
"95": "
|
| 257 |
"96": "SUCCESS",
|
| 258 |
"97": "NON_PARTICIPATING",
|
| 259 |
-
"98": "
|
| 260 |
"99": "NON_PARTICIPATING"
|
| 261 |
},
|
| 262 |
"architectures": [
|
|
|
|
| 1 |
{
|
| 2 |
+
"_name_or_path": "/covalent/.cache/models/tmpetu_1ydyvkd5t8tt",
|
| 3 |
"activation_function": "gelu_new",
|
| 4 |
"all_reduce_scores": {
|
| 5 |
"0": "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": "SUCCESS",
|
| 17 |
"109": "NON_PARTICIPATING",
|
|
|
|
| 20 |
"111": "NON_PARTICIPATING",
|
| 21 |
"112": "NON_PARTICIPATING",
|
| 22 |
"113": "NON_PARTICIPATING",
|
| 23 |
+
"114": "SUCCESS",
|
| 24 |
"115": "SUCCESS",
|
| 25 |
"116": "NON_PARTICIPATING",
|
| 26 |
+
"117": "SUCCESS",
|
| 27 |
"118": "NON_PARTICIPATING",
|
| 28 |
"119": "SUCCESS",
|
| 29 |
+
"12": "SUCCESS",
|
| 30 |
"120": "NON_PARTICIPATING",
|
| 31 |
"121": "SUCCESS",
|
| 32 |
+
"122": "NON_PARTICIPATING",
|
| 33 |
+
"123": "SUCCESS",
|
| 34 |
"124": "SUCCESS",
|
| 35 |
"125": "NON_PARTICIPATING",
|
| 36 |
"126": "NON_PARTICIPATING",
|
| 37 |
"127": "NON_PARTICIPATING",
|
| 38 |
"128": "NON_PARTICIPATING",
|
| 39 |
"129": "SUCCESS",
|
| 40 |
+
"13": "NON_PARTICIPATING",
|
| 41 |
"130": "NON_PARTICIPATING",
|
| 42 |
"131": "NON_PARTICIPATING",
|
| 43 |
"132": "NON_PARTICIPATING",
|
| 44 |
"133": "NON_PARTICIPATING",
|
| 45 |
+
"134": "SUCCESS",
|
| 46 |
+
"135": "NON_PARTICIPATING",
|
| 47 |
"136": "NON_PARTICIPATING",
|
| 48 |
+
"137": "NON_PARTICIPATING",
|
| 49 |
"138": "SUCCESS",
|
| 50 |
"139": "NON_PARTICIPATING",
|
| 51 |
"14": "SUCCESS",
|
| 52 |
"140": "NON_PARTICIPATING",
|
| 53 |
"141": "SUCCESS",
|
| 54 |
"142": "SUCCESS",
|
| 55 |
+
"143": "NON_PARTICIPATING",
|
| 56 |
"144": "NON_PARTICIPATING",
|
| 57 |
"145": "SUCCESS",
|
| 58 |
+
"146": "NON_PARTICIPATING",
|
| 59 |
+
"147": "NON_PARTICIPATING",
|
| 60 |
"148": "NON_PARTICIPATING",
|
| 61 |
"149": "NON_PARTICIPATING",
|
| 62 |
"15": "NON_PARTICIPATING",
|
|
|
|
| 64 |
"151": "SUCCESS",
|
| 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": "SUCCESS",
|
| 73 |
+
"16": "SUCCESS",
|
| 74 |
"160": "NON_PARTICIPATING",
|
| 75 |
"161": "NON_PARTICIPATING",
|
| 76 |
"162": "NON_PARTICIPATING",
|
|
|
|
| 78 |
"164": "SUCCESS",
|
| 79 |
"165": "NON_PARTICIPATING",
|
| 80 |
"166": "NON_PARTICIPATING",
|
| 81 |
+
"167": "NON_PARTICIPATING",
|
| 82 |
+
"168": "SUCCESS",
|
| 83 |
"169": "NON_PARTICIPATING",
|
| 84 |
"17": "NON_PARTICIPATING",
|
| 85 |
"170": "NON_PARTICIPATING",
|
|
|
|
| 94 |
"179": "SUCCESS",
|
| 95 |
"18": "NON_PARTICIPATING",
|
| 96 |
"180": "NON_PARTICIPATING",
|
| 97 |
+
"181": "NON_PARTICIPATING",
|
| 98 |
"182": "NON_PARTICIPATING",
|
| 99 |
"183": "NON_PARTICIPATING",
|
| 100 |
"184": "NON_PARTICIPATING",
|
|
|
|
| 114 |
"197": "NON_PARTICIPATING",
|
| 115 |
"198": "NON_PARTICIPATING",
|
| 116 |
"199": "NON_PARTICIPATING",
|
| 117 |
+
"2": "SUCCESS",
|
| 118 |
"20": "NON_PARTICIPATING",
|
| 119 |
"200": "NON_PARTICIPATING",
|
| 120 |
"201": "SUCCESS",
|
| 121 |
"202": "SUCCESS",
|
| 122 |
"203": "NON_PARTICIPATING",
|
| 123 |
+
"204": "NON_PARTICIPATING",
|
| 124 |
"205": "SUCCESS",
|
| 125 |
"206": "NON_PARTICIPATING",
|
| 126 |
"207": "NON_PARTICIPATING",
|
| 127 |
+
"208": "SUCCESS",
|
| 128 |
+
"209": "SUCCESS",
|
| 129 |
+
"21": "NON_PARTICIPATING",
|
| 130 |
"210": "NON_PARTICIPATING",
|
| 131 |
"211": "NON_PARTICIPATING",
|
| 132 |
"212": "NON_PARTICIPATING",
|
| 133 |
+
"213": "NON_PARTICIPATING",
|
| 134 |
"214": "NON_PARTICIPATING",
|
| 135 |
"215": "NON_PARTICIPATING",
|
| 136 |
+
"216": "NON_PARTICIPATING",
|
| 137 |
+
"217": "NON_PARTICIPATING",
|
| 138 |
"218": "NON_PARTICIPATING",
|
| 139 |
"219": "NON_PARTICIPATING",
|
| 140 |
"22": "SUCCESS",
|
|
|
|
| 143 |
"222": "NON_PARTICIPATING",
|
| 144 |
"223": "NON_PARTICIPATING",
|
| 145 |
"224": "NON_PARTICIPATING",
|
| 146 |
+
"225": "SUCCESS",
|
| 147 |
"226": "NON_PARTICIPATING",
|
| 148 |
"227": "NON_PARTICIPATING",
|
| 149 |
"228": "NON_PARTICIPATING",
|
|
|
|
| 151 |
"23": "SUCCESS",
|
| 152 |
"230": "NON_PARTICIPATING",
|
| 153 |
"231": "NON_PARTICIPATING",
|
| 154 |
+
"232": "SUCCESS",
|
| 155 |
"233": "NON_PARTICIPATING",
|
| 156 |
"234": "NON_PARTICIPATING",
|
| 157 |
+
"235": "SUCCESS",
|
| 158 |
+
"236": "SUCCESS",
|
| 159 |
"237": "NON_PARTICIPATING",
|
| 160 |
"238": "NON_PARTICIPATING",
|
| 161 |
+
"239": "SUCCESS",
|
| 162 |
"24": "NON_PARTICIPATING",
|
| 163 |
"240": "NON_PARTICIPATING",
|
| 164 |
"241": "NON_PARTICIPATING",
|
| 165 |
"242": "NON_PARTICIPATING",
|
| 166 |
"243": "NON_PARTICIPATING",
|
| 167 |
+
"244": "SUCCESS",
|
| 168 |
"245": "NON_PARTICIPATING",
|
| 169 |
"246": "NON_PARTICIPATING",
|
| 170 |
"247": "NON_PARTICIPATING",
|
|
|
|
| 178 |
"254": "NON_PARTICIPATING",
|
| 179 |
"255": "NON_PARTICIPATING",
|
| 180 |
"26": "NON_PARTICIPATING",
|
| 181 |
+
"27": "SUCCESS",
|
| 182 |
"28": "SUCCESS",
|
| 183 |
"29": "NON_PARTICIPATING",
|
| 184 |
"3": "NON_PARTICIPATING",
|
|
|
|
| 206 |
"5": "NON_PARTICIPATING",
|
| 207 |
"50": "SUCCESS",
|
| 208 |
"51": "NON_PARTICIPATING",
|
| 209 |
+
"52": "NON_PARTICIPATING",
|
| 210 |
"53": "NON_PARTICIPATING",
|
| 211 |
"54": "SUCCESS",
|
| 212 |
"55": "SUCCESS",
|
| 213 |
+
"56": "SUCCESS",
|
| 214 |
"57": "NON_PARTICIPATING",
|
| 215 |
"58": "NON_PARTICIPATING",
|
| 216 |
"59": "NON_PARTICIPATING",
|
| 217 |
+
"6": "NON_PARTICIPATING",
|
| 218 |
"60": "NON_PARTICIPATING",
|
| 219 |
"61": "NON_PARTICIPATING",
|
| 220 |
+
"62": "SUCCESS",
|
| 221 |
+
"63": "SUCCESS",
|
| 222 |
"64": "NON_PARTICIPATING",
|
| 223 |
"65": "NON_PARTICIPATING",
|
| 224 |
+
"66": "SUCCESS",
|
| 225 |
+
"67": "SUCCESS",
|
| 226 |
"68": "NON_PARTICIPATING",
|
| 227 |
"69": "NON_PARTICIPATING",
|
| 228 |
"7": "NON_PARTICIPATING",
|
|
|
|
| 233 |
"74": "SUCCESS",
|
| 234 |
"75": "SUCCESS",
|
| 235 |
"76": "NON_PARTICIPATING",
|
| 236 |
+
"77": "SUCCESS",
|
| 237 |
"78": "SUCCESS",
|
| 238 |
+
"79": "SUCCESS",
|
| 239 |
"8": "NON_PARTICIPATING",
|
| 240 |
"80": "NON_PARTICIPATING",
|
| 241 |
"81": "NON_PARTICIPATING",
|
| 242 |
+
"82": "SUCCESS",
|
| 243 |
"83": "NON_PARTICIPATING",
|
| 244 |
+
"84": "SUCCESS",
|
| 245 |
"85": "NON_PARTICIPATING",
|
| 246 |
+
"86": "NON_PARTICIPATING",
|
| 247 |
"87": "NON_PARTICIPATING",
|
| 248 |
"88": "SUCCESS",
|
| 249 |
"89": "NON_PARTICIPATING",
|
| 250 |
+
"9": "NON_PARTICIPATING",
|
| 251 |
"90": "NON_PARTICIPATING",
|
| 252 |
"91": "NON_PARTICIPATING",
|
| 253 |
"92": "NON_PARTICIPATING",
|
| 254 |
+
"93": "NON_PARTICIPATING",
|
| 255 |
"94": "NON_PARTICIPATING",
|
| 256 |
+
"95": "SUCCESS",
|
| 257 |
"96": "SUCCESS",
|
| 258 |
"97": "NON_PARTICIPATING",
|
| 259 |
+
"98": "NON_PARTICIPATING",
|
| 260 |
"99": "NON_PARTICIPATING"
|
| 261 |
},
|
| 262 |
"architectures": [
|
model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1207575528
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2626b9b4e505b54f2d4207d9715c34b4d3f00cb6ddfeaf117befc536af01eff2
|
| 3 |
size 1207575528
|
smash_config.json
CHANGED
|
@@ -28,7 +28,7 @@
|
|
| 28 |
"quant_llm-int8_weight_bits": 8,
|
| 29 |
"max_batch_size": 1,
|
| 30 |
"device": "cuda",
|
| 31 |
-
"cache_dir": "/covalent/.cache/models/
|
| 32 |
"task": "",
|
| 33 |
"save_load_fn": "bitsandbytes",
|
| 34 |
"save_load_fn_args": {}
|
|
|
|
| 28 |
"quant_llm-int8_weight_bits": 8,
|
| 29 |
"max_batch_size": 1,
|
| 30 |
"device": "cuda",
|
| 31 |
+
"cache_dir": "/covalent/.cache/models/tmpetu_1ydy",
|
| 32 |
"task": "",
|
| 33 |
"save_load_fn": "bitsandbytes",
|
| 34 |
"save_load_fn_args": {}
|