286ff9e43765cdc4388bdb53e5e766eef57a365c9000595715fe2d623e63b594
Browse files- config.json +77 -77
- model.safetensors +1 -1
- smash_config.json +1 -1
config.json
CHANGED
@@ -1,10 +1,10 @@
|
|
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": "NON_PARTICIPATING",
|
10 |
"102": "NON_PARTICIPATING",
|
@@ -14,28 +14,28 @@
|
|
14 |
"106": "NON_PARTICIPATING",
|
15 |
"107": "NON_PARTICIPATING",
|
16 |
"108": "NON_PARTICIPATING",
|
17 |
-
"109": "
|
18 |
-
"11": "
|
19 |
"110": "NON_PARTICIPATING",
|
20 |
"111": "NON_PARTICIPATING",
|
21 |
"112": "NON_PARTICIPATING",
|
22 |
"113": "NON_PARTICIPATING",
|
23 |
-
"114": "
|
24 |
"115": "NON_PARTICIPATING",
|
25 |
"116": "NON_PARTICIPATING",
|
26 |
"117": "NON_PARTICIPATING",
|
27 |
-
"118": "
|
28 |
-
"119": "
|
29 |
-
"12": "
|
30 |
"120": "NON_PARTICIPATING",
|
31 |
-
"121": "
|
32 |
"122": "NON_PARTICIPATING",
|
33 |
"123": "SUCCESS",
|
34 |
-
"124": "
|
35 |
"125": "NON_PARTICIPATING",
|
36 |
"126": "NON_PARTICIPATING",
|
37 |
-
"127": "
|
38 |
-
"128": "
|
39 |
"129": "NON_PARTICIPATING",
|
40 |
"13": "NON_PARTICIPATING",
|
41 |
"130": "NON_PARTICIPATING",
|
@@ -48,17 +48,17 @@
|
|
48 |
"137": "NON_PARTICIPATING",
|
49 |
"138": "NON_PARTICIPATING",
|
50 |
"139": "NON_PARTICIPATING",
|
51 |
-
"14": "
|
52 |
"140": "NON_PARTICIPATING",
|
53 |
"141": "NON_PARTICIPATING",
|
54 |
-
"142": "
|
55 |
-
"143": "
|
56 |
"144": "NON_PARTICIPATING",
|
57 |
"145": "NON_PARTICIPATING",
|
58 |
-
"146": "
|
59 |
-
"147": "
|
60 |
"148": "NON_PARTICIPATING",
|
61 |
-
"149": "
|
62 |
"15": "NON_PARTICIPATING",
|
63 |
"150": "NON_PARTICIPATING",
|
64 |
"151": "NON_PARTICIPATING",
|
@@ -70,20 +70,20 @@
|
|
70 |
"157": "NON_PARTICIPATING",
|
71 |
"158": "NON_PARTICIPATING",
|
72 |
"159": "NON_PARTICIPATING",
|
73 |
-
"16": "
|
74 |
-
"160": "
|
75 |
"161": "NON_PARTICIPATING",
|
76 |
"162": "NON_PARTICIPATING",
|
77 |
-
"163": "
|
78 |
"164": "SUCCESS",
|
79 |
"165": "NON_PARTICIPATING",
|
80 |
"166": "NON_PARTICIPATING",
|
81 |
-
"167": "
|
82 |
"168": "NON_PARTICIPATING",
|
83 |
"169": "NON_PARTICIPATING",
|
84 |
"17": "NON_PARTICIPATING",
|
85 |
-
"170": "
|
86 |
-
"171": "
|
87 |
"172": "SUCCESS",
|
88 |
"173": "NON_PARTICIPATING",
|
89 |
"174": "NON_PARTICIPATING",
|
@@ -101,86 +101,86 @@
|
|
101 |
"185": "NON_PARTICIPATING",
|
102 |
"186": "NON_PARTICIPATING",
|
103 |
"187": "NON_PARTICIPATING",
|
104 |
-
"188": "
|
105 |
"189": "NON_PARTICIPATING",
|
106 |
-
"19": "
|
107 |
"190": "NON_PARTICIPATING",
|
108 |
"191": "NON_PARTICIPATING",
|
109 |
"192": "NON_PARTICIPATING",
|
110 |
-
"193": "
|
111 |
"194": "NON_PARTICIPATING",
|
112 |
-
"195": "
|
113 |
"196": "NON_PARTICIPATING",
|
114 |
"197": "NON_PARTICIPATING",
|
115 |
-
"198": "
|
116 |
-
"199": "
|
117 |
"2": "NON_PARTICIPATING",
|
118 |
"20": "NON_PARTICIPATING",
|
119 |
"200": "SUCCESS",
|
120 |
-
"201": "
|
121 |
"202": "SUCCESS",
|
122 |
-
"203": "
|
123 |
"204": "NON_PARTICIPATING",
|
124 |
"205": "NON_PARTICIPATING",
|
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": "SUCCESS",
|
134 |
-
"214": "
|
135 |
"215": "NON_PARTICIPATING",
|
136 |
"216": "NON_PARTICIPATING",
|
137 |
"217": "NON_PARTICIPATING",
|
138 |
"218": "NON_PARTICIPATING",
|
139 |
"219": "NON_PARTICIPATING",
|
140 |
-
"22": "
|
141 |
"220": "SUCCESS",
|
142 |
"221": "NON_PARTICIPATING",
|
143 |
-
"222": "
|
144 |
"223": "NON_PARTICIPATING",
|
145 |
-
"224": "
|
146 |
"225": "NON_PARTICIPATING",
|
147 |
"226": "NON_PARTICIPATING",
|
148 |
-
"227": "
|
149 |
-
"228": "
|
150 |
-
"229": "
|
151 |
"23": "NON_PARTICIPATING",
|
152 |
-
"230": "
|
153 |
-
"231": "
|
154 |
-
"232": "
|
155 |
"233": "NON_PARTICIPATING",
|
156 |
"234": "NON_PARTICIPATING",
|
157 |
"235": "SUCCESS",
|
158 |
"236": "SUCCESS",
|
159 |
-
"237": "
|
160 |
-
"238": "
|
161 |
-
"239": "
|
162 |
"24": "NON_PARTICIPATING",
|
163 |
-
"240": "
|
164 |
-
"241": "
|
165 |
"242": "NON_PARTICIPATING",
|
166 |
"243": "NON_PARTICIPATING",
|
167 |
-
"244": "
|
168 |
-
"245": "
|
169 |
"246": "NON_PARTICIPATING",
|
170 |
"247": "NON_PARTICIPATING",
|
171 |
-
"248": "
|
172 |
-
"249": "
|
173 |
"25": "SUCCESS",
|
174 |
"250": "NON_PARTICIPATING",
|
175 |
"251": "SUCCESS",
|
176 |
"252": "NON_PARTICIPATING",
|
177 |
-
"253": "
|
178 |
"254": "NON_PARTICIPATING",
|
179 |
"255": "NON_PARTICIPATING",
|
180 |
"26": "NON_PARTICIPATING",
|
181 |
"27": "NON_PARTICIPATING",
|
182 |
-
"28": "
|
183 |
-
"29": "
|
184 |
"3": "NON_PARTICIPATING",
|
185 |
"30": "NON_PARTICIPATING",
|
186 |
"31": "NON_PARTICIPATING",
|
@@ -189,10 +189,10 @@
|
|
189 |
"34": "NON_PARTICIPATING",
|
190 |
"35": "NON_PARTICIPATING",
|
191 |
"36": "NON_PARTICIPATING",
|
192 |
-
"37": "
|
193 |
"38": "NON_PARTICIPATING",
|
194 |
"39": "NON_PARTICIPATING",
|
195 |
-
"4": "
|
196 |
"40": "NON_PARTICIPATING",
|
197 |
"41": "NON_PARTICIPATING",
|
198 |
"42": "NON_PARTICIPATING",
|
@@ -202,42 +202,42 @@
|
|
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",
|
209 |
"52": "SUCCESS",
|
210 |
"53": "SUCCESS",
|
211 |
"54": "NON_PARTICIPATING",
|
212 |
-
"55": "
|
213 |
-
"56": "
|
214 |
-
"57": "
|
215 |
-
"58": "
|
216 |
-
"59": "
|
217 |
"6": "NON_PARTICIPATING",
|
218 |
"60": "NON_PARTICIPATING",
|
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": "
|
227 |
-
"69": "
|
228 |
"7": "SUCCESS",
|
229 |
-
"70": "
|
230 |
-
"71": "
|
231 |
"72": "NON_PARTICIPATING",
|
232 |
"73": "NON_PARTICIPATING",
|
233 |
"74": "NON_PARTICIPATING",
|
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": "NON_PARTICIPATING",
|
243 |
"83": "NON_PARTICIPATING",
|
@@ -247,16 +247,16 @@
|
|
247 |
"87": "SUCCESS",
|
248 |
"88": "NON_PARTICIPATING",
|
249 |
"89": "NON_PARTICIPATING",
|
250 |
-
"9": "
|
251 |
-
"90": "
|
252 |
"91": "NON_PARTICIPATING",
|
253 |
"92": "NON_PARTICIPATING",
|
254 |
"93": "NON_PARTICIPATING",
|
255 |
-
"94": "
|
256 |
"95": "NON_PARTICIPATING",
|
257 |
"96": "NON_PARTICIPATING",
|
258 |
"97": "NON_PARTICIPATING",
|
259 |
-
"98": "
|
260 |
"99": "NON_PARTICIPATING"
|
261 |
},
|
262 |
"architectures": [
|
|
|
1 |
{
|
2 |
+
"_name_or_path": "/covalent/.cache/models/tmpbcb54m53kibff9oo",
|
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": "NON_PARTICIPATING",
|
10 |
"102": "NON_PARTICIPATING",
|
|
|
14 |
"106": "NON_PARTICIPATING",
|
15 |
"107": "NON_PARTICIPATING",
|
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": "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",
|
31 |
+
"121": "SUCCESS",
|
32 |
"122": "NON_PARTICIPATING",
|
33 |
"123": "SUCCESS",
|
34 |
+
"124": "NON_PARTICIPATING",
|
35 |
"125": "NON_PARTICIPATING",
|
36 |
"126": "NON_PARTICIPATING",
|
37 |
+
"127": "SUCCESS",
|
38 |
+
"128": "SUCCESS",
|
39 |
"129": "NON_PARTICIPATING",
|
40 |
"13": "NON_PARTICIPATING",
|
41 |
"130": "NON_PARTICIPATING",
|
|
|
48 |
"137": "NON_PARTICIPATING",
|
49 |
"138": "NON_PARTICIPATING",
|
50 |
"139": "NON_PARTICIPATING",
|
51 |
+
"14": "SUCCESS",
|
52 |
"140": "NON_PARTICIPATING",
|
53 |
"141": "NON_PARTICIPATING",
|
54 |
+
"142": "SUCCESS",
|
55 |
+
"143": "SUCCESS",
|
56 |
"144": "NON_PARTICIPATING",
|
57 |
"145": "NON_PARTICIPATING",
|
58 |
+
"146": "NON_PARTICIPATING",
|
59 |
+
"147": "SUCCESS",
|
60 |
"148": "NON_PARTICIPATING",
|
61 |
+
"149": "SUCCESS",
|
62 |
"15": "NON_PARTICIPATING",
|
63 |
"150": "NON_PARTICIPATING",
|
64 |
"151": "NON_PARTICIPATING",
|
|
|
70 |
"157": "NON_PARTICIPATING",
|
71 |
"158": "NON_PARTICIPATING",
|
72 |
"159": "NON_PARTICIPATING",
|
73 |
+
"16": "SUCCESS",
|
74 |
+
"160": "SUCCESS",
|
75 |
"161": "NON_PARTICIPATING",
|
76 |
"162": "NON_PARTICIPATING",
|
77 |
+
"163": "NON_PARTICIPATING",
|
78 |
"164": "SUCCESS",
|
79 |
"165": "NON_PARTICIPATING",
|
80 |
"166": "NON_PARTICIPATING",
|
81 |
+
"167": "SUCCESS",
|
82 |
"168": "NON_PARTICIPATING",
|
83 |
"169": "NON_PARTICIPATING",
|
84 |
"17": "NON_PARTICIPATING",
|
85 |
+
"170": "SUCCESS",
|
86 |
+
"171": "SUCCESS",
|
87 |
"172": "SUCCESS",
|
88 |
"173": "NON_PARTICIPATING",
|
89 |
"174": "NON_PARTICIPATING",
|
|
|
101 |
"185": "NON_PARTICIPATING",
|
102 |
"186": "NON_PARTICIPATING",
|
103 |
"187": "NON_PARTICIPATING",
|
104 |
+
"188": "NON_PARTICIPATING",
|
105 |
"189": "NON_PARTICIPATING",
|
106 |
+
"19": "SUCCESS",
|
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": "SUCCESS",
|
116 |
+
"199": "SUCCESS",
|
117 |
"2": "NON_PARTICIPATING",
|
118 |
"20": "NON_PARTICIPATING",
|
119 |
"200": "SUCCESS",
|
120 |
+
"201": "SUCCESS",
|
121 |
"202": "SUCCESS",
|
122 |
+
"203": "SUCCESS",
|
123 |
"204": "NON_PARTICIPATING",
|
124 |
"205": "NON_PARTICIPATING",
|
125 |
"206": "NON_PARTICIPATING",
|
126 |
"207": "NON_PARTICIPATING",
|
127 |
+
"208": "NON_PARTICIPATING",
|
128 |
+
"209": "NON_PARTICIPATING",
|
129 |
+
"21": "SUCCESS",
|
130 |
"210": "NON_PARTICIPATING",
|
131 |
"211": "NON_PARTICIPATING",
|
132 |
"212": "NON_PARTICIPATING",
|
133 |
"213": "SUCCESS",
|
134 |
+
"214": "SUCCESS",
|
135 |
"215": "NON_PARTICIPATING",
|
136 |
"216": "NON_PARTICIPATING",
|
137 |
"217": "NON_PARTICIPATING",
|
138 |
"218": "NON_PARTICIPATING",
|
139 |
"219": "NON_PARTICIPATING",
|
140 |
+
"22": "SUCCESS",
|
141 |
"220": "SUCCESS",
|
142 |
"221": "NON_PARTICIPATING",
|
143 |
+
"222": "SUCCESS",
|
144 |
"223": "NON_PARTICIPATING",
|
145 |
+
"224": "SUCCESS",
|
146 |
"225": "NON_PARTICIPATING",
|
147 |
"226": "NON_PARTICIPATING",
|
148 |
+
"227": "SUCCESS",
|
149 |
+
"228": "SUCCESS",
|
150 |
+
"229": "NON_PARTICIPATING",
|
151 |
"23": "NON_PARTICIPATING",
|
152 |
+
"230": "SUCCESS",
|
153 |
+
"231": "SUCCESS",
|
154 |
+
"232": "SUCCESS",
|
155 |
"233": "NON_PARTICIPATING",
|
156 |
"234": "NON_PARTICIPATING",
|
157 |
"235": "SUCCESS",
|
158 |
"236": "SUCCESS",
|
159 |
+
"237": "SUCCESS",
|
160 |
+
"238": "SUCCESS",
|
161 |
+
"239": "SUCCESS",
|
162 |
"24": "NON_PARTICIPATING",
|
163 |
+
"240": "SUCCESS",
|
164 |
+
"241": "SUCCESS",
|
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": "SUCCESS",
|
172 |
+
"249": "NON_PARTICIPATING",
|
173 |
"25": "SUCCESS",
|
174 |
"250": "NON_PARTICIPATING",
|
175 |
"251": "SUCCESS",
|
176 |
"252": "NON_PARTICIPATING",
|
177 |
+
"253": "SUCCESS",
|
178 |
"254": "NON_PARTICIPATING",
|
179 |
"255": "NON_PARTICIPATING",
|
180 |
"26": "NON_PARTICIPATING",
|
181 |
"27": "NON_PARTICIPATING",
|
182 |
+
"28": "SUCCESS",
|
183 |
+
"29": "SUCCESS",
|
184 |
"3": "NON_PARTICIPATING",
|
185 |
"30": "NON_PARTICIPATING",
|
186 |
"31": "NON_PARTICIPATING",
|
|
|
189 |
"34": "NON_PARTICIPATING",
|
190 |
"35": "NON_PARTICIPATING",
|
191 |
"36": "NON_PARTICIPATING",
|
192 |
+
"37": "SUCCESS",
|
193 |
"38": "NON_PARTICIPATING",
|
194 |
"39": "NON_PARTICIPATING",
|
195 |
+
"4": "SUCCESS",
|
196 |
"40": "NON_PARTICIPATING",
|
197 |
"41": "NON_PARTICIPATING",
|
198 |
"42": "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",
|
209 |
"52": "SUCCESS",
|
210 |
"53": "SUCCESS",
|
211 |
"54": "NON_PARTICIPATING",
|
212 |
+
"55": "NON_PARTICIPATING",
|
213 |
+
"56": "SUCCESS",
|
214 |
+
"57": "SUCCESS",
|
215 |
+
"58": "NON_PARTICIPATING",
|
216 |
+
"59": "NON_PARTICIPATING",
|
217 |
"6": "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": "NON_PARTICIPATING",
|
225 |
"67": "NON_PARTICIPATING",
|
226 |
+
"68": "SUCCESS",
|
227 |
+
"69": "SUCCESS",
|
228 |
"7": "SUCCESS",
|
229 |
+
"70": "SUCCESS",
|
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": "NON_PARTICIPATING",
|
239 |
"8": "NON_PARTICIPATING",
|
240 |
+
"80": "SUCCESS",
|
241 |
"81": "NON_PARTICIPATING",
|
242 |
"82": "NON_PARTICIPATING",
|
243 |
"83": "NON_PARTICIPATING",
|
|
|
247 |
"87": "SUCCESS",
|
248 |
"88": "NON_PARTICIPATING",
|
249 |
"89": "NON_PARTICIPATING",
|
250 |
+
"9": "SUCCESS",
|
251 |
+
"90": "NON_PARTICIPATING",
|
252 |
"91": "NON_PARTICIPATING",
|
253 |
"92": "NON_PARTICIPATING",
|
254 |
"93": "NON_PARTICIPATING",
|
255 |
+
"94": "SUCCESS",
|
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 sha256:061439ccdf5751e2863fe88b7a4956c1e9edd1b9e2a7adeabdf2249070a6322a
|
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/tmpbcb54m53",
|
32 |
"task": "",
|
33 |
"save_load_fn": "bitsandbytes",
|
34 |
"save_load_fn_args": {}
|