Upload folder using huggingface_hub (#97)
Browse files- 7d80d9d9efc51fa10b78f1559b381ea496537796b429cb171e9f7c2f8a34c5ee (4f468966cfe5386f16282f5bf5675f0a4eb1853f)
- config.json +49 -49
- 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",
|
|
@@ -10,52 +10,52 @@
|
|
| 10 |
"102": "NON_PARTICIPATING",
|
| 11 |
"103": "NON_PARTICIPATING",
|
| 12 |
"104": "NON_PARTICIPATING",
|
| 13 |
-
"105": "
|
| 14 |
"106": "NON_PARTICIPATING",
|
| 15 |
"107": "NON_PARTICIPATING",
|
| 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": "
|
| 24 |
"115": "SUCCESS",
|
| 25 |
"116": "NON_PARTICIPATING",
|
| 26 |
-
"117": "
|
| 27 |
"118": "NON_PARTICIPATING",
|
| 28 |
"119": "NON_PARTICIPATING",
|
| 29 |
-
"12": "
|
| 30 |
"120": "NON_PARTICIPATING",
|
| 31 |
"121": "SUCCESS",
|
| 32 |
-
"122": "
|
| 33 |
"123": "NON_PARTICIPATING",
|
| 34 |
-
"124": "
|
| 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": "NON_PARTICIPATING",
|
| 46 |
-
"135": "
|
| 47 |
"136": "NON_PARTICIPATING",
|
| 48 |
"137": "NON_PARTICIPATING",
|
| 49 |
-
"138": "
|
| 50 |
"139": "NON_PARTICIPATING",
|
| 51 |
-
"14": "
|
| 52 |
"140": "NON_PARTICIPATING",
|
| 53 |
"141": "SUCCESS",
|
| 54 |
-
"142": "
|
| 55 |
"143": "SUCCESS",
|
| 56 |
"144": "NON_PARTICIPATING",
|
| 57 |
"145": "SUCCESS",
|
| 58 |
-
"146": "
|
| 59 |
"147": "SUCCESS",
|
| 60 |
"148": "SUCCESS",
|
| 61 |
"149": "NON_PARTICIPATING",
|
|
@@ -66,11 +66,11 @@
|
|
| 66 |
"153": "NON_PARTICIPATING",
|
| 67 |
"154": "NON_PARTICIPATING",
|
| 68 |
"155": "SUCCESS",
|
| 69 |
-
"156": "
|
| 70 |
"157": "NON_PARTICIPATING",
|
| 71 |
"158": "NON_PARTICIPATING",
|
| 72 |
"159": "NON_PARTICIPATING",
|
| 73 |
-
"16": "
|
| 74 |
"160": "NON_PARTICIPATING",
|
| 75 |
"161": "NON_PARTICIPATING",
|
| 76 |
"162": "NON_PARTICIPATING",
|
|
@@ -78,13 +78,13 @@
|
|
| 78 |
"164": "NON_PARTICIPATING",
|
| 79 |
"165": "NON_PARTICIPATING",
|
| 80 |
"166": "NON_PARTICIPATING",
|
| 81 |
-
"167": "
|
| 82 |
"168": "SUCCESS",
|
| 83 |
"169": "SUCCESS",
|
| 84 |
"17": "NON_PARTICIPATING",
|
| 85 |
"170": "NON_PARTICIPATING",
|
| 86 |
"171": "NON_PARTICIPATING",
|
| 87 |
-
"172": "
|
| 88 |
"173": "NON_PARTICIPATING",
|
| 89 |
"174": "NON_PARTICIPATING",
|
| 90 |
"175": "NON_PARTICIPATING",
|
|
@@ -106,25 +106,25 @@
|
|
| 106 |
"19": "NON_PARTICIPATING",
|
| 107 |
"190": "NON_PARTICIPATING",
|
| 108 |
"191": "NON_PARTICIPATING",
|
| 109 |
-
"192": "
|
| 110 |
"193": "NON_PARTICIPATING",
|
| 111 |
"194": "NON_PARTICIPATING",
|
| 112 |
-
"195": "
|
| 113 |
"196": "NON_PARTICIPATING",
|
| 114 |
"197": "NON_PARTICIPATING",
|
| 115 |
"198": "SUCCESS",
|
| 116 |
"199": "NON_PARTICIPATING",
|
| 117 |
-
"2": "
|
| 118 |
"20": "NON_PARTICIPATING",
|
| 119 |
"200": "NON_PARTICIPATING",
|
| 120 |
-
"201": "
|
| 121 |
"202": "SUCCESS",
|
| 122 |
"203": "NON_PARTICIPATING",
|
| 123 |
-
"204": "
|
| 124 |
-
"205": "
|
| 125 |
"206": "NON_PARTICIPATING",
|
| 126 |
"207": "NON_PARTICIPATING",
|
| 127 |
-
"208": "
|
| 128 |
"209": "SUCCESS",
|
| 129 |
"21": "NON_PARTICIPATING",
|
| 130 |
"210": "NON_PARTICIPATING",
|
|
@@ -134,7 +134,7 @@
|
|
| 134 |
"214": "NON_PARTICIPATING",
|
| 135 |
"215": "NON_PARTICIPATING",
|
| 136 |
"216": "SUCCESS",
|
| 137 |
-
"217": "
|
| 138 |
"218": "NON_PARTICIPATING",
|
| 139 |
"219": "NON_PARTICIPATING",
|
| 140 |
"22": "SUCCESS",
|
|
@@ -142,25 +142,25 @@
|
|
| 142 |
"221": "NON_PARTICIPATING",
|
| 143 |
"222": "NON_PARTICIPATING",
|
| 144 |
"223": "NON_PARTICIPATING",
|
| 145 |
-
"224": "
|
| 146 |
"225": "NON_PARTICIPATING",
|
| 147 |
"226": "NON_PARTICIPATING",
|
| 148 |
-
"227": "
|
| 149 |
-
"228": "
|
| 150 |
"229": "SUCCESS",
|
| 151 |
"23": "SUCCESS",
|
| 152 |
-
"230": "
|
| 153 |
"231": "SUCCESS",
|
| 154 |
"232": "NON_PARTICIPATING",
|
| 155 |
"233": "SUCCESS",
|
| 156 |
-
"234": "
|
| 157 |
"235": "NON_PARTICIPATING",
|
| 158 |
-
"236": "
|
| 159 |
-
"237": "
|
| 160 |
-
"238": "
|
| 161 |
"239": "SUCCESS",
|
| 162 |
"24": "NON_PARTICIPATING",
|
| 163 |
-
"240": "
|
| 164 |
"241": "NON_PARTICIPATING",
|
| 165 |
"242": "NON_PARTICIPATING",
|
| 166 |
"243": "NON_PARTICIPATING",
|
|
@@ -197,16 +197,16 @@
|
|
| 197 |
"41": "NON_PARTICIPATING",
|
| 198 |
"42": "NON_PARTICIPATING",
|
| 199 |
"43": "NON_PARTICIPATING",
|
| 200 |
-
"44": "
|
| 201 |
"45": "SUCCESS",
|
| 202 |
"46": "NON_PARTICIPATING",
|
| 203 |
"47": "NON_PARTICIPATING",
|
| 204 |
"48": "NON_PARTICIPATING",
|
| 205 |
"49": "NON_PARTICIPATING",
|
| 206 |
"5": "NON_PARTICIPATING",
|
| 207 |
-
"50": "
|
| 208 |
"51": "NON_PARTICIPATING",
|
| 209 |
-
"52": "
|
| 210 |
"53": "NON_PARTICIPATING",
|
| 211 |
"54": "NON_PARTICIPATING",
|
| 212 |
"55": "NON_PARTICIPATING",
|
|
@@ -218,26 +218,26 @@
|
|
| 218 |
"60": "NON_PARTICIPATING",
|
| 219 |
"61": "NON_PARTICIPATING",
|
| 220 |
"62": "NON_PARTICIPATING",
|
| 221 |
-
"63": "
|
| 222 |
"64": "NON_PARTICIPATING",
|
| 223 |
-
"65": "
|
| 224 |
-
"66": "
|
| 225 |
-
"67": "
|
| 226 |
"68": "NON_PARTICIPATING",
|
| 227 |
"69": "NON_PARTICIPATING",
|
| 228 |
"7": "SUCCESS",
|
| 229 |
"70": "NON_PARTICIPATING",
|
| 230 |
"71": "NON_PARTICIPATING",
|
| 231 |
"72": "NON_PARTICIPATING",
|
| 232 |
-
"73": "
|
| 233 |
"74": "NON_PARTICIPATING",
|
| 234 |
"75": "NON_PARTICIPATING",
|
| 235 |
"76": "NON_PARTICIPATING",
|
| 236 |
"77": "SUCCESS",
|
| 237 |
-
"78": "
|
| 238 |
"79": "SUCCESS",
|
| 239 |
"8": "NON_PARTICIPATING",
|
| 240 |
-
"80": "
|
| 241 |
"81": "NON_PARTICIPATING",
|
| 242 |
"82": "SUCCESS",
|
| 243 |
"83": "NON_PARTICIPATING",
|
|
@@ -245,18 +245,18 @@
|
|
| 245 |
"85": "NON_PARTICIPATING",
|
| 246 |
"86": "SUCCESS",
|
| 247 |
"87": "NON_PARTICIPATING",
|
| 248 |
-
"88": "
|
| 249 |
"89": "NON_PARTICIPATING",
|
| 250 |
-
"9": "
|
| 251 |
"90": "NON_PARTICIPATING",
|
| 252 |
"91": "NON_PARTICIPATING",
|
| 253 |
"92": "NON_PARTICIPATING",
|
| 254 |
"93": "SUCCESS",
|
| 255 |
"94": "NON_PARTICIPATING",
|
| 256 |
"95": "NON_PARTICIPATING",
|
| 257 |
-
"96": "
|
| 258 |
"97": "NON_PARTICIPATING",
|
| 259 |
-
"98": "
|
| 260 |
"99": "NON_PARTICIPATING"
|
| 261 |
},
|
| 262 |
"architectures": [
|
|
|
|
| 1 |
{
|
| 2 |
+
"_name_or_path": "/covalent/.cache/models/tmp0drue_188gtvz844",
|
| 3 |
"activation_function": "gelu_new",
|
| 4 |
"all_reduce_scores": {
|
| 5 |
"0": "NON_PARTICIPATING",
|
|
|
|
| 10 |
"102": "NON_PARTICIPATING",
|
| 11 |
"103": "NON_PARTICIPATING",
|
| 12 |
"104": "NON_PARTICIPATING",
|
| 13 |
+
"105": "SUCCESS",
|
| 14 |
"106": "NON_PARTICIPATING",
|
| 15 |
"107": "NON_PARTICIPATING",
|
| 16 |
"108": "NON_PARTICIPATING",
|
| 17 |
"109": "NON_PARTICIPATING",
|
| 18 |
"11": "NON_PARTICIPATING",
|
| 19 |
+
"110": "SUCCESS",
|
| 20 |
"111": "NON_PARTICIPATING",
|
| 21 |
"112": "NON_PARTICIPATING",
|
| 22 |
"113": "NON_PARTICIPATING",
|
| 23 |
+
"114": "SUCCESS",
|
| 24 |
"115": "SUCCESS",
|
| 25 |
"116": "NON_PARTICIPATING",
|
| 26 |
+
"117": "NON_PARTICIPATING",
|
| 27 |
"118": "NON_PARTICIPATING",
|
| 28 |
"119": "NON_PARTICIPATING",
|
| 29 |
+
"12": "SUCCESS",
|
| 30 |
"120": "NON_PARTICIPATING",
|
| 31 |
"121": "SUCCESS",
|
| 32 |
+
"122": "NON_PARTICIPATING",
|
| 33 |
"123": "NON_PARTICIPATING",
|
| 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": "NON_PARTICIPATING",
|
| 46 |
+
"135": "SUCCESS",
|
| 47 |
"136": "NON_PARTICIPATING",
|
| 48 |
"137": "NON_PARTICIPATING",
|
| 49 |
+
"138": "NON_PARTICIPATING",
|
| 50 |
"139": "NON_PARTICIPATING",
|
| 51 |
+
"14": "SUCCESS",
|
| 52 |
"140": "NON_PARTICIPATING",
|
| 53 |
"141": "SUCCESS",
|
| 54 |
+
"142": "NON_PARTICIPATING",
|
| 55 |
"143": "SUCCESS",
|
| 56 |
"144": "NON_PARTICIPATING",
|
| 57 |
"145": "SUCCESS",
|
| 58 |
+
"146": "NON_PARTICIPATING",
|
| 59 |
"147": "SUCCESS",
|
| 60 |
"148": "SUCCESS",
|
| 61 |
"149": "NON_PARTICIPATING",
|
|
|
|
| 66 |
"153": "NON_PARTICIPATING",
|
| 67 |
"154": "NON_PARTICIPATING",
|
| 68 |
"155": "SUCCESS",
|
| 69 |
+
"156": "SUCCESS",
|
| 70 |
"157": "NON_PARTICIPATING",
|
| 71 |
"158": "NON_PARTICIPATING",
|
| 72 |
"159": "NON_PARTICIPATING",
|
| 73 |
+
"16": "NON_PARTICIPATING",
|
| 74 |
"160": "NON_PARTICIPATING",
|
| 75 |
"161": "NON_PARTICIPATING",
|
| 76 |
"162": "NON_PARTICIPATING",
|
|
|
|
| 78 |
"164": "NON_PARTICIPATING",
|
| 79 |
"165": "NON_PARTICIPATING",
|
| 80 |
"166": "NON_PARTICIPATING",
|
| 81 |
+
"167": "SUCCESS",
|
| 82 |
"168": "SUCCESS",
|
| 83 |
"169": "SUCCESS",
|
| 84 |
"17": "NON_PARTICIPATING",
|
| 85 |
"170": "NON_PARTICIPATING",
|
| 86 |
"171": "NON_PARTICIPATING",
|
| 87 |
+
"172": "NON_PARTICIPATING",
|
| 88 |
"173": "NON_PARTICIPATING",
|
| 89 |
"174": "NON_PARTICIPATING",
|
| 90 |
"175": "NON_PARTICIPATING",
|
|
|
|
| 106 |
"19": "NON_PARTICIPATING",
|
| 107 |
"190": "NON_PARTICIPATING",
|
| 108 |
"191": "NON_PARTICIPATING",
|
| 109 |
+
"192": "SUCCESS",
|
| 110 |
"193": "NON_PARTICIPATING",
|
| 111 |
"194": "NON_PARTICIPATING",
|
| 112 |
+
"195": "NON_PARTICIPATING",
|
| 113 |
"196": "NON_PARTICIPATING",
|
| 114 |
"197": "NON_PARTICIPATING",
|
| 115 |
"198": "SUCCESS",
|
| 116 |
"199": "NON_PARTICIPATING",
|
| 117 |
+
"2": "NON_PARTICIPATING",
|
| 118 |
"20": "NON_PARTICIPATING",
|
| 119 |
"200": "NON_PARTICIPATING",
|
| 120 |
+
"201": "SUCCESS",
|
| 121 |
"202": "SUCCESS",
|
| 122 |
"203": "NON_PARTICIPATING",
|
| 123 |
+
"204": "SUCCESS",
|
| 124 |
+
"205": "SUCCESS",
|
| 125 |
"206": "NON_PARTICIPATING",
|
| 126 |
"207": "NON_PARTICIPATING",
|
| 127 |
+
"208": "NON_PARTICIPATING",
|
| 128 |
"209": "SUCCESS",
|
| 129 |
"21": "NON_PARTICIPATING",
|
| 130 |
"210": "NON_PARTICIPATING",
|
|
|
|
| 134 |
"214": "NON_PARTICIPATING",
|
| 135 |
"215": "NON_PARTICIPATING",
|
| 136 |
"216": "SUCCESS",
|
| 137 |
+
"217": "NON_PARTICIPATING",
|
| 138 |
"218": "NON_PARTICIPATING",
|
| 139 |
"219": "NON_PARTICIPATING",
|
| 140 |
"22": "SUCCESS",
|
|
|
|
| 142 |
"221": "NON_PARTICIPATING",
|
| 143 |
"222": "NON_PARTICIPATING",
|
| 144 |
"223": "NON_PARTICIPATING",
|
| 145 |
+
"224": "NON_PARTICIPATING",
|
| 146 |
"225": "NON_PARTICIPATING",
|
| 147 |
"226": "NON_PARTICIPATING",
|
| 148 |
+
"227": "SUCCESS",
|
| 149 |
+
"228": "NON_PARTICIPATING",
|
| 150 |
"229": "SUCCESS",
|
| 151 |
"23": "SUCCESS",
|
| 152 |
+
"230": "NON_PARTICIPATING",
|
| 153 |
"231": "SUCCESS",
|
| 154 |
"232": "NON_PARTICIPATING",
|
| 155 |
"233": "SUCCESS",
|
| 156 |
+
"234": "NON_PARTICIPATING",
|
| 157 |
"235": "NON_PARTICIPATING",
|
| 158 |
+
"236": "SUCCESS",
|
| 159 |
+
"237": "SUCCESS",
|
| 160 |
+
"238": "SUCCESS",
|
| 161 |
"239": "SUCCESS",
|
| 162 |
"24": "NON_PARTICIPATING",
|
| 163 |
+
"240": "NON_PARTICIPATING",
|
| 164 |
"241": "NON_PARTICIPATING",
|
| 165 |
"242": "NON_PARTICIPATING",
|
| 166 |
"243": "NON_PARTICIPATING",
|
|
|
|
| 197 |
"41": "NON_PARTICIPATING",
|
| 198 |
"42": "NON_PARTICIPATING",
|
| 199 |
"43": "NON_PARTICIPATING",
|
| 200 |
+
"44": "NON_PARTICIPATING",
|
| 201 |
"45": "SUCCESS",
|
| 202 |
"46": "NON_PARTICIPATING",
|
| 203 |
"47": "NON_PARTICIPATING",
|
| 204 |
"48": "NON_PARTICIPATING",
|
| 205 |
"49": "NON_PARTICIPATING",
|
| 206 |
"5": "NON_PARTICIPATING",
|
| 207 |
+
"50": "NON_PARTICIPATING",
|
| 208 |
"51": "NON_PARTICIPATING",
|
| 209 |
+
"52": "SUCCESS",
|
| 210 |
"53": "NON_PARTICIPATING",
|
| 211 |
"54": "NON_PARTICIPATING",
|
| 212 |
"55": "NON_PARTICIPATING",
|
|
|
|
| 218 |
"60": "NON_PARTICIPATING",
|
| 219 |
"61": "NON_PARTICIPATING",
|
| 220 |
"62": "NON_PARTICIPATING",
|
| 221 |
+
"63": "NON_PARTICIPATING",
|
| 222 |
"64": "NON_PARTICIPATING",
|
| 223 |
+
"65": "SUCCESS",
|
| 224 |
+
"66": "SUCCESS",
|
| 225 |
+
"67": "SUCCESS",
|
| 226 |
"68": "NON_PARTICIPATING",
|
| 227 |
"69": "NON_PARTICIPATING",
|
| 228 |
"7": "SUCCESS",
|
| 229 |
"70": "NON_PARTICIPATING",
|
| 230 |
"71": "NON_PARTICIPATING",
|
| 231 |
"72": "NON_PARTICIPATING",
|
| 232 |
+
"73": "NON_PARTICIPATING",
|
| 233 |
"74": "NON_PARTICIPATING",
|
| 234 |
"75": "NON_PARTICIPATING",
|
| 235 |
"76": "NON_PARTICIPATING",
|
| 236 |
"77": "SUCCESS",
|
| 237 |
+
"78": "SUCCESS",
|
| 238 |
"79": "SUCCESS",
|
| 239 |
"8": "NON_PARTICIPATING",
|
| 240 |
+
"80": "SUCCESS",
|
| 241 |
"81": "NON_PARTICIPATING",
|
| 242 |
"82": "SUCCESS",
|
| 243 |
"83": "NON_PARTICIPATING",
|
|
|
|
| 245 |
"85": "NON_PARTICIPATING",
|
| 246 |
"86": "SUCCESS",
|
| 247 |
"87": "NON_PARTICIPATING",
|
| 248 |
+
"88": "NON_PARTICIPATING",
|
| 249 |
"89": "NON_PARTICIPATING",
|
| 250 |
+
"9": "NON_PARTICIPATING",
|
| 251 |
"90": "NON_PARTICIPATING",
|
| 252 |
"91": "NON_PARTICIPATING",
|
| 253 |
"92": "NON_PARTICIPATING",
|
| 254 |
"93": "SUCCESS",
|
| 255 |
"94": "NON_PARTICIPATING",
|
| 256 |
"95": "NON_PARTICIPATING",
|
| 257 |
+
"96": "NON_PARTICIPATING",
|
| 258 |
"97": "NON_PARTICIPATING",
|
| 259 |
+
"98": "SUCCESS",
|
| 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:efd68bae52b1176d5af5ecda075e4b4e6e80ea0861d99e7744ab20453023d899
|
| 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/tmp0drue_18",
|
| 32 |
"task": "",
|
| 33 |
"save_load_fn": "bitsandbytes",
|
| 34 |
"save_load_fn_args": {}
|