e3ef037d5b85022f1fb31607103e6057f10a6ca39859f71216ee06bf5895a39f
Browse files- config.json +42 -42
- 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",
|
@@ -14,24 +14,24 @@
|
|
14 |
"106": "NON_PARTICIPATING",
|
15 |
"107": "NON_PARTICIPATING",
|
16 |
"108": "NON_PARTICIPATING",
|
17 |
-
"109": "
|
18 |
"11": "NON_PARTICIPATING",
|
19 |
"110": "NON_PARTICIPATING",
|
20 |
"111": "NON_PARTICIPATING",
|
21 |
"112": "NON_PARTICIPATING",
|
22 |
"113": "NON_PARTICIPATING",
|
23 |
-
"114": "
|
24 |
"115": "NON_PARTICIPATING",
|
25 |
-
"116": "
|
26 |
"117": "NON_PARTICIPATING",
|
27 |
"118": "NON_PARTICIPATING",
|
28 |
"119": "NON_PARTICIPATING",
|
29 |
"12": "NON_PARTICIPATING",
|
30 |
"120": "NON_PARTICIPATING",
|
31 |
-
"121": "
|
32 |
"122": "NON_PARTICIPATING",
|
33 |
"123": "NON_PARTICIPATING",
|
34 |
-
"124": "
|
35 |
"125": "NON_PARTICIPATING",
|
36 |
"126": "NON_PARTICIPATING",
|
37 |
"127": "NON_PARTICIPATING",
|
@@ -56,7 +56,7 @@
|
|
56 |
"144": "NON_PARTICIPATING",
|
57 |
"145": "NON_PARTICIPATING",
|
58 |
"146": "NON_PARTICIPATING",
|
59 |
-
"147": "
|
60 |
"148": "NON_PARTICIPATING",
|
61 |
"149": "NON_PARTICIPATING",
|
62 |
"15": "NON_PARTICIPATING",
|
@@ -71,24 +71,24 @@
|
|
71 |
"158": "NON_PARTICIPATING",
|
72 |
"159": "NON_PARTICIPATING",
|
73 |
"16": "NON_PARTICIPATING",
|
74 |
-
"160": "
|
75 |
"161": "NON_PARTICIPATING",
|
76 |
"162": "NON_PARTICIPATING",
|
77 |
-
"163": "
|
78 |
"164": "NON_PARTICIPATING",
|
79 |
"165": "NON_PARTICIPATING",
|
80 |
"166": "NON_PARTICIPATING",
|
81 |
-
"167": "
|
82 |
"168": "NON_PARTICIPATING",
|
83 |
-
"169": "
|
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": "
|
92 |
"177": "NON_PARTICIPATING",
|
93 |
"178": "NON_PARTICIPATING",
|
94 |
"179": "NON_PARTICIPATING",
|
@@ -104,22 +104,22 @@
|
|
104 |
"188": "NON_PARTICIPATING",
|
105 |
"189": "NON_PARTICIPATING",
|
106 |
"19": "NON_PARTICIPATING",
|
107 |
-
"190": "
|
108 |
"191": "NON_PARTICIPATING",
|
109 |
"192": "NON_PARTICIPATING",
|
110 |
"193": "NON_PARTICIPATING",
|
111 |
-
"194": "
|
112 |
"195": "NON_PARTICIPATING",
|
113 |
-
"196": "
|
114 |
"197": "NON_PARTICIPATING",
|
115 |
-
"198": "
|
116 |
"199": "NON_PARTICIPATING",
|
117 |
"2": "NON_PARTICIPATING",
|
118 |
"20": "NON_PARTICIPATING",
|
119 |
"200": "NON_PARTICIPATING",
|
120 |
-
"201": "
|
121 |
"202": "NON_PARTICIPATING",
|
122 |
-
"203": "
|
123 |
"204": "NON_PARTICIPATING",
|
124 |
"205": "NON_PARTICIPATING",
|
125 |
"206": "NON_PARTICIPATING",
|
@@ -127,34 +127,34 @@
|
|
127 |
"208": "NON_PARTICIPATING",
|
128 |
"209": "NON_PARTICIPATING",
|
129 |
"21": "NON_PARTICIPATING",
|
130 |
-
"210": "
|
131 |
-
"211": "
|
132 |
"212": "NON_PARTICIPATING",
|
133 |
"213": "NON_PARTICIPATING",
|
134 |
-
"214": "
|
135 |
"215": "NON_PARTICIPATING",
|
136 |
"216": "NON_PARTICIPATING",
|
137 |
"217": "NON_PARTICIPATING",
|
138 |
"218": "NON_PARTICIPATING",
|
139 |
"219": "NON_PARTICIPATING",
|
140 |
"22": "NON_PARTICIPATING",
|
141 |
-
"220": "
|
142 |
-
"221": "
|
143 |
-
"222": "
|
144 |
"223": "NON_PARTICIPATING",
|
145 |
"224": "NON_PARTICIPATING",
|
146 |
"225": "NON_PARTICIPATING",
|
147 |
"226": "NON_PARTICIPATING",
|
148 |
-
"227": "
|
149 |
-
"228": "
|
150 |
"229": "NON_PARTICIPATING",
|
151 |
"23": "NON_PARTICIPATING",
|
152 |
"230": "NON_PARTICIPATING",
|
153 |
"231": "NON_PARTICIPATING",
|
154 |
-
"232": "
|
155 |
"233": "NON_PARTICIPATING",
|
156 |
"234": "NON_PARTICIPATING",
|
157 |
-
"235": "
|
158 |
"236": "NON_PARTICIPATING",
|
159 |
"237": "NON_PARTICIPATING",
|
160 |
"238": "NON_PARTICIPATING",
|
@@ -164,17 +164,17 @@
|
|
164 |
"241": "NON_PARTICIPATING",
|
165 |
"242": "NON_PARTICIPATING",
|
166 |
"243": "NON_PARTICIPATING",
|
167 |
-
"244": "
|
168 |
-
"245": "
|
169 |
"246": "NON_PARTICIPATING",
|
170 |
"247": "NON_PARTICIPATING",
|
171 |
"248": "NON_PARTICIPATING",
|
172 |
"249": "NON_PARTICIPATING",
|
173 |
"25": "SUCCESS",
|
174 |
-
"250": "
|
175 |
"251": "NON_PARTICIPATING",
|
176 |
"252": "NON_PARTICIPATING",
|
177 |
-
"253": "
|
178 |
"254": "NON_PARTICIPATING",
|
179 |
"255": "NON_PARTICIPATING",
|
180 |
"26": "NON_PARTICIPATING",
|
@@ -188,7 +188,7 @@
|
|
188 |
"33": "NON_PARTICIPATING",
|
189 |
"34": "NON_PARTICIPATING",
|
190 |
"35": "NON_PARTICIPATING",
|
191 |
-
"36": "
|
192 |
"37": "NON_PARTICIPATING",
|
193 |
"38": "NON_PARTICIPATING",
|
194 |
"39": "NON_PARTICIPATING",
|
@@ -202,7 +202,7 @@
|
|
202 |
"46": "NON_PARTICIPATING",
|
203 |
"47": "NON_PARTICIPATING",
|
204 |
"48": "NON_PARTICIPATING",
|
205 |
-
"49": "
|
206 |
"5": "NON_PARTICIPATING",
|
207 |
"50": "NON_PARTICIPATING",
|
208 |
"51": "NON_PARTICIPATING",
|
@@ -211,16 +211,16 @@
|
|
211 |
"54": "NON_PARTICIPATING",
|
212 |
"55": "NON_PARTICIPATING",
|
213 |
"56": "NON_PARTICIPATING",
|
214 |
-
"57": "
|
215 |
"58": "NON_PARTICIPATING",
|
216 |
"59": "NON_PARTICIPATING",
|
217 |
"6": "SUCCESS",
|
218 |
-
"60": "
|
219 |
"61": "NON_PARTICIPATING",
|
220 |
"62": "NON_PARTICIPATING",
|
221 |
"63": "NON_PARTICIPATING",
|
222 |
"64": "NON_PARTICIPATING",
|
223 |
-
"65": "
|
224 |
"66": "NON_PARTICIPATING",
|
225 |
"67": "NON_PARTICIPATING",
|
226 |
"68": "NON_PARTICIPATING",
|
@@ -229,7 +229,7 @@
|
|
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",
|
@@ -237,11 +237,11 @@
|
|
237 |
"78": "NON_PARTICIPATING",
|
238 |
"79": "NON_PARTICIPATING",
|
239 |
"8": "NON_PARTICIPATING",
|
240 |
-
"80": "
|
241 |
"81": "NON_PARTICIPATING",
|
242 |
-
"82": "
|
243 |
"83": "NON_PARTICIPATING",
|
244 |
-
"84": "
|
245 |
"85": "NON_PARTICIPATING",
|
246 |
"86": "NON_PARTICIPATING",
|
247 |
"87": "NON_PARTICIPATING",
|
|
|
1 |
{
|
2 |
+
"_name_or_path": "/covalent/.cache/models/tmp91xcf65v663wygaa",
|
3 |
"activation_function": "gelu_new",
|
4 |
"all_reduce_scores": {
|
5 |
"0": "NON_PARTICIPATING",
|
|
|
14 |
"106": "NON_PARTICIPATING",
|
15 |
"107": "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": "NON_PARTICIPATING",
|
28 |
"119": "NON_PARTICIPATING",
|
29 |
"12": "NON_PARTICIPATING",
|
30 |
"120": "NON_PARTICIPATING",
|
31 |
+
"121": "SUCCESS",
|
32 |
"122": "NON_PARTICIPATING",
|
33 |
"123": "NON_PARTICIPATING",
|
34 |
+
"124": "NON_PARTICIPATING",
|
35 |
"125": "NON_PARTICIPATING",
|
36 |
"126": "NON_PARTICIPATING",
|
37 |
"127": "NON_PARTICIPATING",
|
|
|
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": "NON_PARTICIPATING",
|
|
|
71 |
"158": "NON_PARTICIPATING",
|
72 |
"159": "NON_PARTICIPATING",
|
73 |
"16": "NON_PARTICIPATING",
|
74 |
+
"160": "SUCCESS",
|
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": "NON_PARTICIPATING",
|
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",
|
91 |
+
"176": "NON_PARTICIPATING",
|
92 |
"177": "NON_PARTICIPATING",
|
93 |
"178": "NON_PARTICIPATING",
|
94 |
"179": "NON_PARTICIPATING",
|
|
|
104 |
"188": "NON_PARTICIPATING",
|
105 |
"189": "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": "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": "NON_PARTICIPATING",
|
121 |
"202": "NON_PARTICIPATING",
|
122 |
+
"203": "SUCCESS",
|
123 |
"204": "NON_PARTICIPATING",
|
124 |
"205": "NON_PARTICIPATING",
|
125 |
"206": "NON_PARTICIPATING",
|
|
|
127 |
"208": "NON_PARTICIPATING",
|
128 |
"209": "NON_PARTICIPATING",
|
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": "NON_PARTICIPATING",
|
141 |
+
"220": "NON_PARTICIPATING",
|
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": "NON_PARTICIPATING",
|
149 |
+
"228": "NON_PARTICIPATING",
|
150 |
"229": "NON_PARTICIPATING",
|
151 |
"23": "NON_PARTICIPATING",
|
152 |
"230": "NON_PARTICIPATING",
|
153 |
"231": "NON_PARTICIPATING",
|
154 |
+
"232": "SUCCESS",
|
155 |
"233": "NON_PARTICIPATING",
|
156 |
"234": "NON_PARTICIPATING",
|
157 |
+
"235": "SUCCESS",
|
158 |
"236": "NON_PARTICIPATING",
|
159 |
"237": "NON_PARTICIPATING",
|
160 |
"238": "NON_PARTICIPATING",
|
|
|
164 |
"241": "NON_PARTICIPATING",
|
165 |
"242": "NON_PARTICIPATING",
|
166 |
"243": "NON_PARTICIPATING",
|
167 |
+
"244": "NON_PARTICIPATING",
|
168 |
+
"245": "SUCCESS",
|
169 |
"246": "NON_PARTICIPATING",
|
170 |
"247": "NON_PARTICIPATING",
|
171 |
"248": "NON_PARTICIPATING",
|
172 |
"249": "NON_PARTICIPATING",
|
173 |
"25": "SUCCESS",
|
174 |
+
"250": "NON_PARTICIPATING",
|
175 |
"251": "NON_PARTICIPATING",
|
176 |
"252": "NON_PARTICIPATING",
|
177 |
+
"253": "SUCCESS",
|
178 |
"254": "NON_PARTICIPATING",
|
179 |
"255": "NON_PARTICIPATING",
|
180 |
"26": "NON_PARTICIPATING",
|
|
|
188 |
"33": "NON_PARTICIPATING",
|
189 |
"34": "NON_PARTICIPATING",
|
190 |
"35": "NON_PARTICIPATING",
|
191 |
+
"36": "NON_PARTICIPATING",
|
192 |
"37": "NON_PARTICIPATING",
|
193 |
"38": "NON_PARTICIPATING",
|
194 |
"39": "NON_PARTICIPATING",
|
|
|
202 |
"46": "NON_PARTICIPATING",
|
203 |
"47": "NON_PARTICIPATING",
|
204 |
"48": "NON_PARTICIPATING",
|
205 |
+
"49": "SUCCESS",
|
206 |
"5": "NON_PARTICIPATING",
|
207 |
"50": "NON_PARTICIPATING",
|
208 |
"51": "NON_PARTICIPATING",
|
|
|
211 |
"54": "NON_PARTICIPATING",
|
212 |
"55": "NON_PARTICIPATING",
|
213 |
"56": "NON_PARTICIPATING",
|
214 |
+
"57": "SUCCESS",
|
215 |
"58": "NON_PARTICIPATING",
|
216 |
"59": "NON_PARTICIPATING",
|
217 |
"6": "SUCCESS",
|
218 |
+
"60": "SUCCESS",
|
219 |
"61": "NON_PARTICIPATING",
|
220 |
"62": "NON_PARTICIPATING",
|
221 |
"63": "NON_PARTICIPATING",
|
222 |
"64": "NON_PARTICIPATING",
|
223 |
+
"65": "SUCCESS",
|
224 |
"66": "NON_PARTICIPATING",
|
225 |
"67": "NON_PARTICIPATING",
|
226 |
"68": "NON_PARTICIPATING",
|
|
|
229 |
"70": "NON_PARTICIPATING",
|
230 |
"71": "NON_PARTICIPATING",
|
231 |
"72": "NON_PARTICIPATING",
|
232 |
+
"73": "SUCCESS",
|
233 |
"74": "NON_PARTICIPATING",
|
234 |
"75": "NON_PARTICIPATING",
|
235 |
"76": "NON_PARTICIPATING",
|
|
|
237 |
"78": "NON_PARTICIPATING",
|
238 |
"79": "NON_PARTICIPATING",
|
239 |
"8": "NON_PARTICIPATING",
|
240 |
+
"80": "NON_PARTICIPATING",
|
241 |
"81": "NON_PARTICIPATING",
|
242 |
+
"82": "NON_PARTICIPATING",
|
243 |
"83": "NON_PARTICIPATING",
|
244 |
+
"84": "NON_PARTICIPATING",
|
245 |
"85": "NON_PARTICIPATING",
|
246 |
"86": "NON_PARTICIPATING",
|
247 |
"87": "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 sha256:099168c4dd3b08916e40d7b4e3557e32631fadc2269da6bd191c8f0a4f63dd15
|
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/tmp91xcf65v",
|
32 |
"task": "",
|
33 |
"save_load_fn": "bitsandbytes",
|
34 |
"save_load_fn_args": {}
|