88039044a82fa0da5685e6ec1e49f96f452cabd4bdfe3c08f63ab5f87812af70
Browse files- config.json +51 -51
- model.safetensors +1 -1
- smash_config.json +1 -1
config.json
CHANGED
|
@@ -1,12 +1,12 @@
|
|
| 1 |
{
|
| 2 |
-
"_name_or_path": "/covalent/.cache/models/
|
| 3 |
"activation_function": "gelu_new",
|
| 4 |
"all_reduce_scores": {
|
| 5 |
"0": "NON_PARTICIPATING",
|
| 6 |
"1": "NON_PARTICIPATING",
|
| 7 |
-
"10": "
|
| 8 |
"100": "NON_PARTICIPATING",
|
| 9 |
-
"101": "
|
| 10 |
"102": "NON_PARTICIPATING",
|
| 11 |
"103": "NON_PARTICIPATING",
|
| 12 |
"104": "NON_PARTICIPATING",
|
|
@@ -16,15 +16,15 @@
|
|
| 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": "SUCCESS",
|
| 24 |
"115": "NON_PARTICIPATING",
|
| 25 |
"116": "NON_PARTICIPATING",
|
| 26 |
-
"117": "
|
| 27 |
-
"118": "
|
| 28 |
"119": "NON_PARTICIPATING",
|
| 29 |
"12": "SUCCESS",
|
| 30 |
"120": "NON_PARTICIPATING",
|
|
@@ -43,10 +43,10 @@
|
|
| 43 |
"132": "NON_PARTICIPATING",
|
| 44 |
"133": "NON_PARTICIPATING",
|
| 45 |
"134": "SUCCESS",
|
| 46 |
-
"135": "
|
| 47 |
"136": "NON_PARTICIPATING",
|
| 48 |
"137": "SUCCESS",
|
| 49 |
-
"138": "
|
| 50 |
"139": "NON_PARTICIPATING",
|
| 51 |
"14": "NON_PARTICIPATING",
|
| 52 |
"140": "NON_PARTICIPATING",
|
|
@@ -57,41 +57,41 @@
|
|
| 57 |
"145": "NON_PARTICIPATING",
|
| 58 |
"146": "NON_PARTICIPATING",
|
| 59 |
"147": "NON_PARTICIPATING",
|
| 60 |
-
"148": "
|
| 61 |
"149": "NON_PARTICIPATING",
|
| 62 |
"15": "NON_PARTICIPATING",
|
| 63 |
"150": "NON_PARTICIPATING",
|
| 64 |
-
"151": "
|
| 65 |
"152": "NON_PARTICIPATING",
|
| 66 |
"153": "NON_PARTICIPATING",
|
| 67 |
"154": "SUCCESS",
|
| 68 |
-
"155": "
|
| 69 |
-
"156": "
|
| 70 |
"157": "NON_PARTICIPATING",
|
| 71 |
-
"158": "
|
| 72 |
-
"159": "
|
| 73 |
"16": "NON_PARTICIPATING",
|
| 74 |
"160": "NON_PARTICIPATING",
|
| 75 |
"161": "NON_PARTICIPATING",
|
| 76 |
"162": "NON_PARTICIPATING",
|
| 77 |
"163": "NON_PARTICIPATING",
|
| 78 |
-
"164": "
|
| 79 |
"165": "NON_PARTICIPATING",
|
| 80 |
"166": "NON_PARTICIPATING",
|
| 81 |
-
"167": "
|
| 82 |
-
"168": "
|
| 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",
|
| 91 |
"176": "NON_PARTICIPATING",
|
| 92 |
"177": "NON_PARTICIPATING",
|
| 93 |
"178": "SUCCESS",
|
| 94 |
-
"179": "
|
| 95 |
"18": "NON_PARTICIPATING",
|
| 96 |
"180": "NON_PARTICIPATING",
|
| 97 |
"181": "NON_PARTICIPATING",
|
|
@@ -106,10 +106,10 @@
|
|
| 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": "NON_PARTICIPATING",
|
|
@@ -117,8 +117,8 @@
|
|
| 117 |
"2": "SUCCESS",
|
| 118 |
"20": "NON_PARTICIPATING",
|
| 119 |
"200": "NON_PARTICIPATING",
|
| 120 |
-
"201": "
|
| 121 |
-
"202": "
|
| 122 |
"203": "NON_PARTICIPATING",
|
| 123 |
"204": "SUCCESS",
|
| 124 |
"205": "SUCCESS",
|
|
@@ -130,10 +130,10 @@
|
|
| 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": "NON_PARTICIPATING",
|
| 138 |
"218": "NON_PARTICIPATING",
|
| 139 |
"219": "NON_PARTICIPATING",
|
|
@@ -142,15 +142,15 @@
|
|
| 142 |
"221": "NON_PARTICIPATING",
|
| 143 |
"222": "NON_PARTICIPATING",
|
| 144 |
"223": "NON_PARTICIPATING",
|
| 145 |
-
"224": "
|
| 146 |
-
"225": "
|
| 147 |
"226": "NON_PARTICIPATING",
|
| 148 |
-
"227": "
|
| 149 |
-
"228": "
|
| 150 |
-
"229": "
|
| 151 |
"23": "NON_PARTICIPATING",
|
| 152 |
-
"230": "
|
| 153 |
-
"231": "
|
| 154 |
"232": "NON_PARTICIPATING",
|
| 155 |
"233": "NON_PARTICIPATING",
|
| 156 |
"234": "NON_PARTICIPATING",
|
|
@@ -160,7 +160,7 @@
|
|
| 160 |
"238": "SUCCESS",
|
| 161 |
"239": "SUCCESS",
|
| 162 |
"24": "NON_PARTICIPATING",
|
| 163 |
-
"240": "
|
| 164 |
"241": "NON_PARTICIPATING",
|
| 165 |
"242": "NON_PARTICIPATING",
|
| 166 |
"243": "NON_PARTICIPATING",
|
|
@@ -173,12 +173,12 @@
|
|
| 173 |
"25": "SUCCESS",
|
| 174 |
"250": "NON_PARTICIPATING",
|
| 175 |
"251": "NON_PARTICIPATING",
|
| 176 |
-
"252": "
|
| 177 |
"253": "NON_PARTICIPATING",
|
| 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",
|
|
@@ -191,7 +191,7 @@
|
|
| 191 |
"36": "NON_PARTICIPATING",
|
| 192 |
"37": "NON_PARTICIPATING",
|
| 193 |
"38": "NON_PARTICIPATING",
|
| 194 |
-
"39": "
|
| 195 |
"4": "NON_PARTICIPATING",
|
| 196 |
"40": "NON_PARTICIPATING",
|
| 197 |
"41": "NON_PARTICIPATING",
|
|
@@ -209,39 +209,39 @@
|
|
| 209 |
"52": "NON_PARTICIPATING",
|
| 210 |
"53": "NON_PARTICIPATING",
|
| 211 |
"54": "SUCCESS",
|
| 212 |
-
"55": "
|
| 213 |
-
"56": "
|
| 214 |
"57": "NON_PARTICIPATING",
|
| 215 |
-
"58": "
|
| 216 |
"59": "NON_PARTICIPATING",
|
| 217 |
"6": "SUCCESS",
|
| 218 |
"60": "NON_PARTICIPATING",
|
| 219 |
"61": "NON_PARTICIPATING",
|
| 220 |
-
"62": "
|
| 221 |
-
"63": "
|
| 222 |
"64": "NON_PARTICIPATING",
|
| 223 |
"65": "SUCCESS",
|
| 224 |
"66": "SUCCESS",
|
| 225 |
-
"67": "
|
| 226 |
"68": "NON_PARTICIPATING",
|
| 227 |
"69": "NON_PARTICIPATING",
|
| 228 |
-
"7": "
|
| 229 |
"70": "NON_PARTICIPATING",
|
| 230 |
"71": "NON_PARTICIPATING",
|
| 231 |
"72": "NON_PARTICIPATING",
|
| 232 |
"73": "SUCCESS",
|
| 233 |
-
"74": "
|
| 234 |
"75": "NON_PARTICIPATING",
|
| 235 |
"76": "NON_PARTICIPATING",
|
| 236 |
-
"77": "
|
| 237 |
-
"78": "
|
| 238 |
"79": "NON_PARTICIPATING",
|
| 239 |
"8": "NON_PARTICIPATING",
|
| 240 |
-
"80": "
|
| 241 |
"81": "NON_PARTICIPATING",
|
| 242 |
"82": "SUCCESS",
|
| 243 |
"83": "NON_PARTICIPATING",
|
| 244 |
-
"84": "
|
| 245 |
"85": "NON_PARTICIPATING",
|
| 246 |
"86": "NON_PARTICIPATING",
|
| 247 |
"87": "NON_PARTICIPATING",
|
|
@@ -251,10 +251,10 @@
|
|
| 251 |
"90": "NON_PARTICIPATING",
|
| 252 |
"91": "NON_PARTICIPATING",
|
| 253 |
"92": "NON_PARTICIPATING",
|
| 254 |
-
"93": "
|
| 255 |
"94": "NON_PARTICIPATING",
|
| 256 |
-
"95": "
|
| 257 |
-
"96": "
|
| 258 |
"97": "NON_PARTICIPATING",
|
| 259 |
"98": "NON_PARTICIPATING",
|
| 260 |
"99": "NON_PARTICIPATING"
|
|
|
|
| 1 |
{
|
| 2 |
+
"_name_or_path": "/covalent/.cache/models/tmp_dl0fyrgili90iwp",
|
| 3 |
"activation_function": "gelu_new",
|
| 4 |
"all_reduce_scores": {
|
| 5 |
"0": "NON_PARTICIPATING",
|
| 6 |
"1": "NON_PARTICIPATING",
|
| 7 |
+
"10": "SUCCESS",
|
| 8 |
"100": "NON_PARTICIPATING",
|
| 9 |
+
"101": "SUCCESS",
|
| 10 |
"102": "NON_PARTICIPATING",
|
| 11 |
"103": "NON_PARTICIPATING",
|
| 12 |
"104": "NON_PARTICIPATING",
|
|
|
|
| 16 |
"108": "NON_PARTICIPATING",
|
| 17 |
"109": "NON_PARTICIPATING",
|
| 18 |
"11": "NON_PARTICIPATING",
|
| 19 |
+
"110": "NON_PARTICIPATING",
|
| 20 |
"111": "NON_PARTICIPATING",
|
| 21 |
"112": "NON_PARTICIPATING",
|
| 22 |
"113": "NON_PARTICIPATING",
|
| 23 |
"114": "SUCCESS",
|
| 24 |
"115": "NON_PARTICIPATING",
|
| 25 |
"116": "NON_PARTICIPATING",
|
| 26 |
+
"117": "NON_PARTICIPATING",
|
| 27 |
+
"118": "SUCCESS",
|
| 28 |
"119": "NON_PARTICIPATING",
|
| 29 |
"12": "SUCCESS",
|
| 30 |
"120": "NON_PARTICIPATING",
|
|
|
|
| 43 |
"132": "NON_PARTICIPATING",
|
| 44 |
"133": "NON_PARTICIPATING",
|
| 45 |
"134": "SUCCESS",
|
| 46 |
+
"135": "NON_PARTICIPATING",
|
| 47 |
"136": "NON_PARTICIPATING",
|
| 48 |
"137": "SUCCESS",
|
| 49 |
+
"138": "SUCCESS",
|
| 50 |
"139": "NON_PARTICIPATING",
|
| 51 |
"14": "NON_PARTICIPATING",
|
| 52 |
"140": "NON_PARTICIPATING",
|
|
|
|
| 57 |
"145": "NON_PARTICIPATING",
|
| 58 |
"146": "NON_PARTICIPATING",
|
| 59 |
"147": "NON_PARTICIPATING",
|
| 60 |
+
"148": "SUCCESS",
|
| 61 |
"149": "NON_PARTICIPATING",
|
| 62 |
"15": "NON_PARTICIPATING",
|
| 63 |
"150": "NON_PARTICIPATING",
|
| 64 |
+
"151": "SUCCESS",
|
| 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": "SUCCESS",
|
| 72 |
+
"159": "SUCCESS",
|
| 73 |
"16": "NON_PARTICIPATING",
|
| 74 |
"160": "NON_PARTICIPATING",
|
| 75 |
"161": "NON_PARTICIPATING",
|
| 76 |
"162": "NON_PARTICIPATING",
|
| 77 |
"163": "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": "SUCCESS",
|
| 88 |
"173": "NON_PARTICIPATING",
|
| 89 |
"174": "NON_PARTICIPATING",
|
| 90 |
"175": "NON_PARTICIPATING",
|
| 91 |
"176": "NON_PARTICIPATING",
|
| 92 |
"177": "NON_PARTICIPATING",
|
| 93 |
"178": "SUCCESS",
|
| 94 |
+
"179": "SUCCESS",
|
| 95 |
"18": "NON_PARTICIPATING",
|
| 96 |
"180": "NON_PARTICIPATING",
|
| 97 |
"181": "NON_PARTICIPATING",
|
|
|
|
| 106 |
"19": "NON_PARTICIPATING",
|
| 107 |
"190": "NON_PARTICIPATING",
|
| 108 |
"191": "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",
|
|
|
|
| 117 |
"2": "SUCCESS",
|
| 118 |
"20": "NON_PARTICIPATING",
|
| 119 |
"200": "NON_PARTICIPATING",
|
| 120 |
+
"201": "SUCCESS",
|
| 121 |
+
"202": "NON_PARTICIPATING",
|
| 122 |
"203": "NON_PARTICIPATING",
|
| 123 |
"204": "SUCCESS",
|
| 124 |
"205": "SUCCESS",
|
|
|
|
| 130 |
"210": "NON_PARTICIPATING",
|
| 131 |
"211": "NON_PARTICIPATING",
|
| 132 |
"212": "NON_PARTICIPATING",
|
| 133 |
+
"213": "SUCCESS",
|
| 134 |
"214": "NON_PARTICIPATING",
|
| 135 |
"215": "NON_PARTICIPATING",
|
| 136 |
+
"216": "SUCCESS",
|
| 137 |
"217": "NON_PARTICIPATING",
|
| 138 |
"218": "NON_PARTICIPATING",
|
| 139 |
"219": "NON_PARTICIPATING",
|
|
|
|
| 142 |
"221": "NON_PARTICIPATING",
|
| 143 |
"222": "NON_PARTICIPATING",
|
| 144 |
"223": "NON_PARTICIPATING",
|
| 145 |
+
"224": "SUCCESS",
|
| 146 |
+
"225": "SUCCESS",
|
| 147 |
"226": "NON_PARTICIPATING",
|
| 148 |
+
"227": "NON_PARTICIPATING",
|
| 149 |
+
"228": "SUCCESS",
|
| 150 |
+
"229": "SUCCESS",
|
| 151 |
"23": "NON_PARTICIPATING",
|
| 152 |
+
"230": "NON_PARTICIPATING",
|
| 153 |
+
"231": "NON_PARTICIPATING",
|
| 154 |
"232": "NON_PARTICIPATING",
|
| 155 |
"233": "NON_PARTICIPATING",
|
| 156 |
"234": "NON_PARTICIPATING",
|
|
|
|
| 160 |
"238": "SUCCESS",
|
| 161 |
"239": "SUCCESS",
|
| 162 |
"24": "NON_PARTICIPATING",
|
| 163 |
+
"240": "SUCCESS",
|
| 164 |
"241": "NON_PARTICIPATING",
|
| 165 |
"242": "NON_PARTICIPATING",
|
| 166 |
"243": "NON_PARTICIPATING",
|
|
|
|
| 173 |
"25": "SUCCESS",
|
| 174 |
"250": "NON_PARTICIPATING",
|
| 175 |
"251": "NON_PARTICIPATING",
|
| 176 |
+
"252": "SUCCESS",
|
| 177 |
"253": "NON_PARTICIPATING",
|
| 178 |
"254": "NON_PARTICIPATING",
|
| 179 |
"255": "NON_PARTICIPATING",
|
| 180 |
"26": "NON_PARTICIPATING",
|
| 181 |
+
"27": "NON_PARTICIPATING",
|
| 182 |
"28": "SUCCESS",
|
| 183 |
"29": "NON_PARTICIPATING",
|
| 184 |
"3": "NON_PARTICIPATING",
|
|
|
|
| 191 |
"36": "NON_PARTICIPATING",
|
| 192 |
"37": "NON_PARTICIPATING",
|
| 193 |
"38": "NON_PARTICIPATING",
|
| 194 |
+
"39": "SUCCESS",
|
| 195 |
"4": "NON_PARTICIPATING",
|
| 196 |
"40": "NON_PARTICIPATING",
|
| 197 |
"41": "NON_PARTICIPATING",
|
|
|
|
| 209 |
"52": "NON_PARTICIPATING",
|
| 210 |
"53": "NON_PARTICIPATING",
|
| 211 |
"54": "SUCCESS",
|
| 212 |
+
"55": "SUCCESS",
|
| 213 |
+
"56": "NON_PARTICIPATING",
|
| 214 |
"57": "NON_PARTICIPATING",
|
| 215 |
+
"58": "SUCCESS",
|
| 216 |
"59": "NON_PARTICIPATING",
|
| 217 |
"6": "SUCCESS",
|
| 218 |
"60": "NON_PARTICIPATING",
|
| 219 |
"61": "NON_PARTICIPATING",
|
| 220 |
+
"62": "NON_PARTICIPATING",
|
| 221 |
+
"63": "SUCCESS",
|
| 222 |
"64": "NON_PARTICIPATING",
|
| 223 |
"65": "SUCCESS",
|
| 224 |
"66": "SUCCESS",
|
| 225 |
+
"67": "NON_PARTICIPATING",
|
| 226 |
"68": "NON_PARTICIPATING",
|
| 227 |
"69": "NON_PARTICIPATING",
|
| 228 |
+
"7": "NON_PARTICIPATING",
|
| 229 |
"70": "NON_PARTICIPATING",
|
| 230 |
"71": "NON_PARTICIPATING",
|
| 231 |
"72": "NON_PARTICIPATING",
|
| 232 |
"73": "SUCCESS",
|
| 233 |
+
"74": "SUCCESS",
|
| 234 |
"75": "NON_PARTICIPATING",
|
| 235 |
"76": "NON_PARTICIPATING",
|
| 236 |
+
"77": "NON_PARTICIPATING",
|
| 237 |
+
"78": "NON_PARTICIPATING",
|
| 238 |
"79": "NON_PARTICIPATING",
|
| 239 |
"8": "NON_PARTICIPATING",
|
| 240 |
+
"80": "SUCCESS",
|
| 241 |
"81": "NON_PARTICIPATING",
|
| 242 |
"82": "SUCCESS",
|
| 243 |
"83": "NON_PARTICIPATING",
|
| 244 |
+
"84": "NON_PARTICIPATING",
|
| 245 |
"85": "NON_PARTICIPATING",
|
| 246 |
"86": "NON_PARTICIPATING",
|
| 247 |
"87": "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": "NON_PARTICIPATING",
|
| 258 |
"97": "NON_PARTICIPATING",
|
| 259 |
"98": "NON_PARTICIPATING",
|
| 260 |
"99": "NON_PARTICIPATING"
|
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 sha256:89bda62ff2de4bc85270cf0af995afc7ccac5378a4dc3509894270b332450697
|
| 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/tmp_dl0fyrg",
|
| 32 |
"task": "",
|
| 33 |
"save_load_fn": "bitsandbytes",
|
| 34 |
"save_load_fn_args": {}
|