Upload folder using huggingface_hub (#233)
Browse files- ebeb1a0a047b6581cb1fdb10dadfe561c2dbd67e3573583ba930a07c2677e327 (bf94beaf8351e7384a26298e7d37f91a02c9c137)
- config.json +53 -53
- model.safetensors +1 -1
- smash_config.json +1 -1
config.json
CHANGED
@@ -1,41 +1,41 @@
|
|
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": "NON_PARTICIPATING",
|
8 |
"100": "NON_PARTICIPATING",
|
9 |
-
"101": "
|
10 |
"102": "NON_PARTICIPATING",
|
11 |
"103": "NON_PARTICIPATING",
|
12 |
"104": "NON_PARTICIPATING",
|
13 |
"105": "NON_PARTICIPATING",
|
14 |
-
"106": "
|
15 |
-
"107": "
|
16 |
"108": "NON_PARTICIPATING",
|
17 |
-
"109": "
|
18 |
-
"11": "
|
19 |
"110": "NON_PARTICIPATING",
|
20 |
-
"111": "
|
21 |
"112": "NON_PARTICIPATING",
|
22 |
"113": "NON_PARTICIPATING",
|
23 |
"114": "NON_PARTICIPATING",
|
24 |
"115": "NON_PARTICIPATING",
|
25 |
-
"116": "
|
26 |
"117": "NON_PARTICIPATING",
|
27 |
-
"118": "
|
28 |
-
"119": "
|
29 |
"12": "NON_PARTICIPATING",
|
30 |
"120": "NON_PARTICIPATING",
|
31 |
"121": "NON_PARTICIPATING",
|
32 |
-
"122": "
|
33 |
"123": "NON_PARTICIPATING",
|
34 |
"124": "NON_PARTICIPATING",
|
35 |
"125": "NON_PARTICIPATING",
|
36 |
"126": "NON_PARTICIPATING",
|
37 |
"127": "NON_PARTICIPATING",
|
38 |
-
"128": "
|
39 |
"129": "NON_PARTICIPATING",
|
40 |
"13": "NON_PARTICIPATING",
|
41 |
"130": "NON_PARTICIPATING",
|
@@ -59,8 +59,8 @@
|
|
59 |
"147": "NON_PARTICIPATING",
|
60 |
"148": "NON_PARTICIPATING",
|
61 |
"149": "NON_PARTICIPATING",
|
62 |
-
"15": "
|
63 |
-
"150": "
|
64 |
"151": "NON_PARTICIPATING",
|
65 |
"152": "NON_PARTICIPATING",
|
66 |
"153": "NON_PARTICIPATING",
|
@@ -70,23 +70,23 @@
|
|
70 |
"157": "SUCCESS",
|
71 |
"158": "NON_PARTICIPATING",
|
72 |
"159": "NON_PARTICIPATING",
|
73 |
-
"16": "
|
74 |
"160": "NON_PARTICIPATING",
|
75 |
"161": "NON_PARTICIPATING",
|
76 |
-
"162": "
|
77 |
"163": "NON_PARTICIPATING",
|
78 |
-
"164": "
|
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": "
|
87 |
-
"172": "
|
88 |
"173": "NON_PARTICIPATING",
|
89 |
-
"174": "
|
90 |
"175": "NON_PARTICIPATING",
|
91 |
"176": "NON_PARTICIPATING",
|
92 |
"177": "NON_PARTICIPATING",
|
@@ -95,17 +95,17 @@
|
|
95 |
"18": "NON_PARTICIPATING",
|
96 |
"180": "NON_PARTICIPATING",
|
97 |
"181": "NON_PARTICIPATING",
|
98 |
-
"182": "
|
99 |
"183": "SUCCESS",
|
100 |
-
"184": "
|
101 |
"185": "NON_PARTICIPATING",
|
102 |
-
"186": "
|
103 |
"187": "NON_PARTICIPATING",
|
104 |
"188": "SUCCESS",
|
105 |
"189": "SUCCESS",
|
106 |
"19": "NON_PARTICIPATING",
|
107 |
"190": "NON_PARTICIPATING",
|
108 |
-
"191": "
|
109 |
"192": "NON_PARTICIPATING",
|
110 |
"193": "SUCCESS",
|
111 |
"194": "SUCCESS",
|
@@ -113,10 +113,10 @@
|
|
113 |
"196": "SUCCESS",
|
114 |
"197": "NON_PARTICIPATING",
|
115 |
"198": "NON_PARTICIPATING",
|
116 |
-
"199": "
|
117 |
"2": "NON_PARTICIPATING",
|
118 |
"20": "NON_PARTICIPATING",
|
119 |
-
"200": "
|
120 |
"201": "NON_PARTICIPATING",
|
121 |
"202": "NON_PARTICIPATING",
|
122 |
"203": "NON_PARTICIPATING",
|
@@ -127,10 +127,10 @@
|
|
127 |
"208": "NON_PARTICIPATING",
|
128 |
"209": "NON_PARTICIPATING",
|
129 |
"21": "NON_PARTICIPATING",
|
130 |
-
"210": "
|
131 |
-
"211": "
|
132 |
-
"212": "
|
133 |
-
"213": "
|
134 |
"214": "NON_PARTICIPATING",
|
135 |
"215": "NON_PARTICIPATING",
|
136 |
"216": "NON_PARTICIPATING",
|
@@ -138,26 +138,26 @@
|
|
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": "SUCCESS",
|
148 |
"227": "NON_PARTICIPATING",
|
149 |
"228": "NON_PARTICIPATING",
|
150 |
-
"229": "
|
151 |
"23": "NON_PARTICIPATING",
|
152 |
-
"230": "
|
153 |
"231": "NON_PARTICIPATING",
|
154 |
"232": "NON_PARTICIPATING",
|
155 |
-
"233": "
|
156 |
"234": "NON_PARTICIPATING",
|
157 |
"235": "NON_PARTICIPATING",
|
158 |
"236": "NON_PARTICIPATING",
|
159 |
"237": "NON_PARTICIPATING",
|
160 |
-
"238": "
|
161 |
"239": "NON_PARTICIPATING",
|
162 |
"24": "NON_PARTICIPATING",
|
163 |
"240": "NON_PARTICIPATING",
|
@@ -169,7 +169,7 @@
|
|
169 |
"246": "NON_PARTICIPATING",
|
170 |
"247": "NON_PARTICIPATING",
|
171 |
"248": "NON_PARTICIPATING",
|
172 |
-
"249": "
|
173 |
"25": "SUCCESS",
|
174 |
"250": "NON_PARTICIPATING",
|
175 |
"251": "NON_PARTICIPATING",
|
@@ -181,9 +181,9 @@
|
|
181 |
"27": "NON_PARTICIPATING",
|
182 |
"28": "NON_PARTICIPATING",
|
183 |
"29": "NON_PARTICIPATING",
|
184 |
-
"3": "
|
185 |
-
"30": "
|
186 |
-
"31": "
|
187 |
"32": "NON_PARTICIPATING",
|
188 |
"33": "NON_PARTICIPATING",
|
189 |
"34": "SUCCESS",
|
@@ -213,7 +213,7 @@
|
|
213 |
"56": "NON_PARTICIPATING",
|
214 |
"57": "NON_PARTICIPATING",
|
215 |
"58": "SUCCESS",
|
216 |
-
"59": "
|
217 |
"6": "NON_PARTICIPATING",
|
218 |
"60": "NON_PARTICIPATING",
|
219 |
"61": "NON_PARTICIPATING",
|
@@ -225,12 +225,12 @@
|
|
225 |
"67": "NON_PARTICIPATING",
|
226 |
"68": "NON_PARTICIPATING",
|
227 |
"69": "NON_PARTICIPATING",
|
228 |
-
"7": "
|
229 |
"70": "NON_PARTICIPATING",
|
230 |
"71": "SUCCESS",
|
231 |
-
"72": "
|
232 |
"73": "NON_PARTICIPATING",
|
233 |
-
"74": "
|
234 |
"75": "NON_PARTICIPATING",
|
235 |
"76": "NON_PARTICIPATING",
|
236 |
"77": "NON_PARTICIPATING",
|
@@ -242,22 +242,22 @@
|
|
242 |
"82": "NON_PARTICIPATING",
|
243 |
"83": "SUCCESS",
|
244 |
"84": "NON_PARTICIPATING",
|
245 |
-
"85": "
|
246 |
"86": "NON_PARTICIPATING",
|
247 |
"87": "NON_PARTICIPATING",
|
248 |
"88": "NON_PARTICIPATING",
|
249 |
"89": "SUCCESS",
|
250 |
-
"9": "
|
251 |
-
"90": "
|
252 |
"91": "NON_PARTICIPATING",
|
253 |
"92": "NON_PARTICIPATING",
|
254 |
"93": "NON_PARTICIPATING",
|
255 |
-
"94": "
|
256 |
"95": "NON_PARTICIPATING",
|
257 |
-
"96": "
|
258 |
"97": "NON_PARTICIPATING",
|
259 |
"98": "NON_PARTICIPATING",
|
260 |
-
"99": "
|
261 |
},
|
262 |
"architectures": [
|
263 |
"GPTOptim"
|
|
|
1 |
{
|
2 |
+
"_name_or_path": "/covalent/.cache/models/tmppwwozur_0py35ys1",
|
3 |
"activation_function": "gelu_new",
|
4 |
"all_reduce_scores": {
|
5 |
"0": "NON_PARTICIPATING",
|
6 |
"1": "NON_PARTICIPATING",
|
7 |
"10": "NON_PARTICIPATING",
|
8 |
"100": "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": "SUCCESS",
|
16 |
"108": "NON_PARTICIPATING",
|
17 |
+
"109": "SUCCESS",
|
18 |
+
"11": "SUCCESS",
|
19 |
"110": "NON_PARTICIPATING",
|
20 |
+
"111": "SUCCESS",
|
21 |
"112": "NON_PARTICIPATING",
|
22 |
"113": "NON_PARTICIPATING",
|
23 |
"114": "NON_PARTICIPATING",
|
24 |
"115": "NON_PARTICIPATING",
|
25 |
+
"116": "NON_PARTICIPATING",
|
26 |
"117": "NON_PARTICIPATING",
|
27 |
+
"118": "SUCCESS",
|
28 |
+
"119": "SUCCESS",
|
29 |
"12": "NON_PARTICIPATING",
|
30 |
"120": "NON_PARTICIPATING",
|
31 |
"121": "NON_PARTICIPATING",
|
32 |
+
"122": "NON_PARTICIPATING",
|
33 |
"123": "NON_PARTICIPATING",
|
34 |
"124": "NON_PARTICIPATING",
|
35 |
"125": "NON_PARTICIPATING",
|
36 |
"126": "NON_PARTICIPATING",
|
37 |
"127": "NON_PARTICIPATING",
|
38 |
+
"128": "NON_PARTICIPATING",
|
39 |
"129": "NON_PARTICIPATING",
|
40 |
"13": "NON_PARTICIPATING",
|
41 |
"130": "NON_PARTICIPATING",
|
|
|
59 |
"147": "NON_PARTICIPATING",
|
60 |
"148": "NON_PARTICIPATING",
|
61 |
"149": "NON_PARTICIPATING",
|
62 |
+
"15": "SUCCESS",
|
63 |
+
"150": "SUCCESS",
|
64 |
"151": "NON_PARTICIPATING",
|
65 |
"152": "NON_PARTICIPATING",
|
66 |
"153": "NON_PARTICIPATING",
|
|
|
70 |
"157": "SUCCESS",
|
71 |
"158": "NON_PARTICIPATING",
|
72 |
"159": "NON_PARTICIPATING",
|
73 |
+
"16": "SUCCESS",
|
74 |
"160": "NON_PARTICIPATING",
|
75 |
"161": "NON_PARTICIPATING",
|
76 |
+
"162": "SUCCESS",
|
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": "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",
|
|
|
95 |
"18": "NON_PARTICIPATING",
|
96 |
"180": "NON_PARTICIPATING",
|
97 |
"181": "NON_PARTICIPATING",
|
98 |
+
"182": "SUCCESS",
|
99 |
"183": "SUCCESS",
|
100 |
+
"184": "SUCCESS",
|
101 |
"185": "NON_PARTICIPATING",
|
102 |
+
"186": "NON_PARTICIPATING",
|
103 |
"187": "NON_PARTICIPATING",
|
104 |
"188": "SUCCESS",
|
105 |
"189": "SUCCESS",
|
106 |
"19": "NON_PARTICIPATING",
|
107 |
"190": "NON_PARTICIPATING",
|
108 |
+
"191": "NON_PARTICIPATING",
|
109 |
"192": "NON_PARTICIPATING",
|
110 |
"193": "SUCCESS",
|
111 |
"194": "SUCCESS",
|
|
|
113 |
"196": "SUCCESS",
|
114 |
"197": "NON_PARTICIPATING",
|
115 |
"198": "NON_PARTICIPATING",
|
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": "NON_PARTICIPATING",
|
|
|
127 |
"208": "NON_PARTICIPATING",
|
128 |
"209": "NON_PARTICIPATING",
|
129 |
"21": "NON_PARTICIPATING",
|
130 |
+
"210": "SUCCESS",
|
131 |
+
"211": "SUCCESS",
|
132 |
+
"212": "SUCCESS",
|
133 |
+
"213": "SUCCESS",
|
134 |
"214": "NON_PARTICIPATING",
|
135 |
"215": "NON_PARTICIPATING",
|
136 |
"216": "NON_PARTICIPATING",
|
|
|
138 |
"218": "NON_PARTICIPATING",
|
139 |
"219": "NON_PARTICIPATING",
|
140 |
"22": "NON_PARTICIPATING",
|
141 |
+
"220": "SUCCESS",
|
142 |
+
"221": "SUCCESS",
|
143 |
+
"222": "NON_PARTICIPATING",
|
144 |
"223": "NON_PARTICIPATING",
|
145 |
"224": "NON_PARTICIPATING",
|
146 |
"225": "NON_PARTICIPATING",
|
147 |
"226": "SUCCESS",
|
148 |
"227": "NON_PARTICIPATING",
|
149 |
"228": "NON_PARTICIPATING",
|
150 |
+
"229": "SUCCESS",
|
151 |
"23": "NON_PARTICIPATING",
|
152 |
+
"230": "SUCCESS",
|
153 |
"231": "NON_PARTICIPATING",
|
154 |
"232": "NON_PARTICIPATING",
|
155 |
+
"233": "NON_PARTICIPATING",
|
156 |
"234": "NON_PARTICIPATING",
|
157 |
"235": "NON_PARTICIPATING",
|
158 |
"236": "NON_PARTICIPATING",
|
159 |
"237": "NON_PARTICIPATING",
|
160 |
+
"238": "SUCCESS",
|
161 |
"239": "NON_PARTICIPATING",
|
162 |
"24": "NON_PARTICIPATING",
|
163 |
"240": "NON_PARTICIPATING",
|
|
|
169 |
"246": "NON_PARTICIPATING",
|
170 |
"247": "NON_PARTICIPATING",
|
171 |
"248": "NON_PARTICIPATING",
|
172 |
+
"249": "SUCCESS",
|
173 |
"25": "SUCCESS",
|
174 |
"250": "NON_PARTICIPATING",
|
175 |
"251": "NON_PARTICIPATING",
|
|
|
181 |
"27": "NON_PARTICIPATING",
|
182 |
"28": "NON_PARTICIPATING",
|
183 |
"29": "NON_PARTICIPATING",
|
184 |
+
"3": "SUCCESS",
|
185 |
+
"30": "SUCCESS",
|
186 |
+
"31": "SUCCESS",
|
187 |
"32": "NON_PARTICIPATING",
|
188 |
"33": "NON_PARTICIPATING",
|
189 |
"34": "SUCCESS",
|
|
|
213 |
"56": "NON_PARTICIPATING",
|
214 |
"57": "NON_PARTICIPATING",
|
215 |
"58": "SUCCESS",
|
216 |
+
"59": "SUCCESS",
|
217 |
"6": "NON_PARTICIPATING",
|
218 |
"60": "NON_PARTICIPATING",
|
219 |
"61": "NON_PARTICIPATING",
|
|
|
225 |
"67": "NON_PARTICIPATING",
|
226 |
"68": "NON_PARTICIPATING",
|
227 |
"69": "NON_PARTICIPATING",
|
228 |
+
"7": "SUCCESS",
|
229 |
"70": "NON_PARTICIPATING",
|
230 |
"71": "SUCCESS",
|
231 |
+
"72": "SUCCESS",
|
232 |
"73": "NON_PARTICIPATING",
|
233 |
+
"74": "SUCCESS",
|
234 |
"75": "NON_PARTICIPATING",
|
235 |
"76": "NON_PARTICIPATING",
|
236 |
"77": "NON_PARTICIPATING",
|
|
|
242 |
"82": "NON_PARTICIPATING",
|
243 |
"83": "SUCCESS",
|
244 |
"84": "NON_PARTICIPATING",
|
245 |
+
"85": "SUCCESS",
|
246 |
"86": "NON_PARTICIPATING",
|
247 |
"87": "NON_PARTICIPATING",
|
248 |
"88": "NON_PARTICIPATING",
|
249 |
"89": "SUCCESS",
|
250 |
+
"9": "SUCCESS",
|
251 |
+
"90": "SUCCESS",
|
252 |
"91": "NON_PARTICIPATING",
|
253 |
"92": "NON_PARTICIPATING",
|
254 |
"93": "NON_PARTICIPATING",
|
255 |
+
"94": "SUCCESS",
|
256 |
"95": "NON_PARTICIPATING",
|
257 |
+
"96": "SUCCESS",
|
258 |
"97": "NON_PARTICIPATING",
|
259 |
"98": "NON_PARTICIPATING",
|
260 |
+
"99": "SUCCESS"
|
261 |
},
|
262 |
"architectures": [
|
263 |
"GPTOptim"
|
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:cb86ce0caf81f719ddd5e6d23cfaeb234526387f5428281aa5010d8670b25500
|
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/tmppwwozur_",
|
32 |
"task": "",
|
33 |
"save_load_fn": "bitsandbytes",
|
34 |
"save_load_fn_args": {}
|