Run 4. Outer Step 84. Inner Step 0. Peers 82.
Browse files- config.json +55 -55
- inner_optimizer.pt +1 -1
- model.safetensors +1 -1
- outer_optimizer.pt +1 -1
config.json
CHANGED
@@ -1,15 +1,15 @@
|
|
1 |
{
|
2 |
-
"_name_or_path": "
|
3 |
"activation_function": "gelu_new",
|
4 |
"all_reduce_scores": {
|
5 |
"0": "NON_PARTICIPATING",
|
6 |
-
"1": "
|
7 |
"10": "NON_PARTICIPATING",
|
8 |
"100": "NON_PARTICIPATING",
|
9 |
"101": "NON_PARTICIPATING",
|
10 |
"102": "NON_PARTICIPATING",
|
11 |
"103": "NON_PARTICIPATING",
|
12 |
-
"104": "
|
13 |
"105": "NON_PARTICIPATING",
|
14 |
"106": "NON_PARTICIPATING",
|
15 |
"107": "NON_PARTICIPATING",
|
@@ -24,7 +24,7 @@
|
|
24 |
"115": "NON_PARTICIPATING",
|
25 |
"116": "NON_PARTICIPATING",
|
26 |
"117": "NON_PARTICIPATING",
|
27 |
-
"118": "
|
28 |
"119": "NON_PARTICIPATING",
|
29 |
"12": "SUCCESS",
|
30 |
"120": "NON_PARTICIPATING",
|
@@ -38,7 +38,7 @@
|
|
38 |
"128": "NON_PARTICIPATING",
|
39 |
"129": "SUCCESS",
|
40 |
"13": "SUCCESS",
|
41 |
-
"130": "
|
42 |
"131": "NON_PARTICIPATING",
|
43 |
"132": "NON_PARTICIPATING",
|
44 |
"133": "NON_PARTICIPATING",
|
@@ -48,7 +48,7 @@
|
|
48 |
"137": "NON_PARTICIPATING",
|
49 |
"138": "NON_PARTICIPATING",
|
50 |
"139": "NON_PARTICIPATING",
|
51 |
-
"14": "
|
52 |
"140": "NON_PARTICIPATING",
|
53 |
"141": "NON_PARTICIPATING",
|
54 |
"142": "NON_PARTICIPATING",
|
@@ -56,43 +56,43 @@
|
|
56 |
"144": "NON_PARTICIPATING",
|
57 |
"145": "NON_PARTICIPATING",
|
58 |
"146": "NON_PARTICIPATING",
|
59 |
-
"147": "
|
60 |
-
"148": "
|
61 |
"149": "NON_PARTICIPATING",
|
62 |
-
"15": "
|
63 |
"150": "NON_PARTICIPATING",
|
64 |
"151": "NON_PARTICIPATING",
|
65 |
-
"152": "
|
66 |
"153": "NON_PARTICIPATING",
|
67 |
"154": "NON_PARTICIPATING",
|
68 |
"155": "NON_PARTICIPATING",
|
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",
|
77 |
"163": "NON_PARTICIPATING",
|
78 |
-
"164": "
|
79 |
"165": "NON_PARTICIPATING",
|
80 |
"166": "NON_PARTICIPATING",
|
81 |
"167": "NON_PARTICIPATING",
|
82 |
"168": "NON_PARTICIPATING",
|
83 |
"169": "NON_PARTICIPATING",
|
84 |
-
"17": "
|
85 |
"170": "NON_PARTICIPATING",
|
86 |
"171": "NON_PARTICIPATING",
|
87 |
"172": "NON_PARTICIPATING",
|
88 |
-
"173": "
|
89 |
"174": "NON_PARTICIPATING",
|
90 |
"175": "NON_PARTICIPATING",
|
91 |
"176": "NON_PARTICIPATING",
|
92 |
"177": "NON_PARTICIPATING",
|
93 |
"178": "NON_PARTICIPATING",
|
94 |
"179": "NON_PARTICIPATING",
|
95 |
-
"18": "
|
96 |
"180": "NON_PARTICIPATING",
|
97 |
"181": "NON_PARTICIPATING",
|
98 |
"182": "NON_PARTICIPATING",
|
@@ -101,7 +101,7 @@
|
|
101 |
"185": "NON_PARTICIPATING",
|
102 |
"186": "NON_PARTICIPATING",
|
103 |
"187": "NON_PARTICIPATING",
|
104 |
-
"188": "
|
105 |
"189": "NON_PARTICIPATING",
|
106 |
"19": "SUCCESS",
|
107 |
"190": "NON_PARTICIPATING",
|
@@ -110,15 +110,15 @@
|
|
110 |
"193": "NON_PARTICIPATING",
|
111 |
"194": "NON_PARTICIPATING",
|
112 |
"195": "NON_PARTICIPATING",
|
113 |
-
"196": "
|
114 |
"197": "NON_PARTICIPATING",
|
115 |
"198": "NON_PARTICIPATING",
|
116 |
"199": "NON_PARTICIPATING",
|
117 |
-
"2": "
|
118 |
"20": "SUCCESS",
|
119 |
"200": "NON_PARTICIPATING",
|
120 |
"201": "NON_PARTICIPATING",
|
121 |
-
"202": "
|
122 |
"203": "NON_PARTICIPATING",
|
123 |
"204": "NON_PARTICIPATING",
|
124 |
"205": "NON_PARTICIPATING",
|
@@ -136,8 +136,8 @@
|
|
136 |
"216": "NON_PARTICIPATING",
|
137 |
"217": "SUCCESS",
|
138 |
"218": "NON_PARTICIPATING",
|
139 |
-
"219": "
|
140 |
-
"22": "
|
141 |
"220": "NON_PARTICIPATING",
|
142 |
"221": "NON_PARTICIPATING",
|
143 |
"222": "NON_PARTICIPATING",
|
@@ -155,11 +155,11 @@
|
|
155 |
"233": "NON_PARTICIPATING",
|
156 |
"234": "NON_PARTICIPATING",
|
157 |
"235": "NON_PARTICIPATING",
|
158 |
-
"236": "
|
159 |
"237": "NON_PARTICIPATING",
|
160 |
"238": "NON_PARTICIPATING",
|
161 |
"239": "NON_PARTICIPATING",
|
162 |
-
"24": "
|
163 |
"240": "NON_PARTICIPATING",
|
164 |
"241": "NON_PARTICIPATING",
|
165 |
"242": "NON_PARTICIPATING",
|
@@ -177,59 +177,59 @@
|
|
177 |
"253": "NON_PARTICIPATING",
|
178 |
"254": "NON_PARTICIPATING",
|
179 |
"255": "NON_PARTICIPATING",
|
180 |
-
"26": "
|
181 |
"27": "SUCCESS",
|
182 |
"28": "SUCCESS",
|
183 |
-
"29": "
|
184 |
"3": "NON_PARTICIPATING",
|
185 |
"30": "NON_PARTICIPATING",
|
186 |
-
"31": "
|
187 |
"32": "NON_PARTICIPATING",
|
188 |
"33": "NON_PARTICIPATING",
|
189 |
"34": "SUCCESS",
|
190 |
-
"35": "
|
191 |
"36": "NON_PARTICIPATING",
|
192 |
"37": "NON_PARTICIPATING",
|
193 |
"38": "SUCCESS",
|
194 |
-
"39": "
|
195 |
-
"4": "
|
196 |
"40": "SUCCESS",
|
197 |
-
"41": "
|
198 |
"42": "SUCCESS",
|
199 |
"43": "NON_PARTICIPATING",
|
200 |
"44": "SUCCESS",
|
201 |
-
"45": "
|
202 |
-
"46": "
|
203 |
"47": "NON_PARTICIPATING",
|
204 |
"48": "NON_PARTICIPATING",
|
205 |
-
"49": "
|
206 |
"5": "SUCCESS",
|
207 |
"50": "NON_PARTICIPATING",
|
208 |
"51": "SUCCESS",
|
209 |
"52": "SUCCESS",
|
210 |
"53": "NON_PARTICIPATING",
|
211 |
-
"54": "
|
212 |
-
"55": "
|
213 |
-
"56": "
|
214 |
"57": "NON_PARTICIPATING",
|
215 |
"58": "NON_PARTICIPATING",
|
216 |
"59": "SUCCESS",
|
217 |
"6": "NON_PARTICIPATING",
|
218 |
"60": "SUCCESS",
|
219 |
-
"61": "
|
220 |
-
"62": "
|
221 |
"63": "SUCCESS",
|
222 |
"64": "SUCCESS",
|
223 |
-
"65": "
|
224 |
"66": "NON_PARTICIPATING",
|
225 |
-
"67": "
|
226 |
"68": "NON_PARTICIPATING",
|
227 |
"69": "NON_PARTICIPATING",
|
228 |
-
"7": "
|
229 |
"70": "NON_PARTICIPATING",
|
230 |
"71": "SUCCESS",
|
231 |
-
"72": "
|
232 |
-
"73": "
|
233 |
"74": "NON_PARTICIPATING",
|
234 |
"75": "NON_PARTICIPATING",
|
235 |
"76": "NON_PARTICIPATING",
|
@@ -237,26 +237,26 @@
|
|
237 |
"78": "NON_PARTICIPATING",
|
238 |
"79": "NON_PARTICIPATING",
|
239 |
"8": "SUCCESS",
|
240 |
-
"80": "
|
241 |
"81": "NON_PARTICIPATING",
|
242 |
"82": "NON_PARTICIPATING",
|
243 |
-
"83": "
|
244 |
-
"84": "
|
245 |
"85": "SUCCESS",
|
246 |
"86": "NON_PARTICIPATING",
|
247 |
-
"87": "
|
248 |
-
"88": "
|
249 |
-
"89": "
|
250 |
"9": "SUCCESS",
|
251 |
-
"90": "
|
252 |
"91": "SUCCESS",
|
253 |
"92": "NON_PARTICIPATING",
|
254 |
"93": "NON_PARTICIPATING",
|
255 |
-
"94": "
|
256 |
"95": "SUCCESS",
|
257 |
"96": "NON_PARTICIPATING",
|
258 |
-
"97": "
|
259 |
-
"98": "
|
260 |
"99": "NON_PARTICIPATING"
|
261 |
},
|
262 |
"architectures": [
|
@@ -275,7 +275,7 @@
|
|
275 |
"initializer_range": 0.02,
|
276 |
"inner_step": 0,
|
277 |
"inner_steps": 0,
|
278 |
-
"last_allreduce_block":
|
279 |
"layer_norm_epsilon": 1e-05,
|
280 |
"model_type": "gpt_optimized",
|
281 |
"n_embd": 1280,
|
|
|
1 |
{
|
2 |
+
"_name_or_path": "distributed/optimized-gpt2-1b",
|
3 |
"activation_function": "gelu_new",
|
4 |
"all_reduce_scores": {
|
5 |
"0": "NON_PARTICIPATING",
|
6 |
+
"1": "SUCCESS",
|
7 |
"10": "NON_PARTICIPATING",
|
8 |
"100": "NON_PARTICIPATING",
|
9 |
"101": "NON_PARTICIPATING",
|
10 |
"102": "NON_PARTICIPATING",
|
11 |
"103": "NON_PARTICIPATING",
|
12 |
+
"104": "SUCCESS",
|
13 |
"105": "NON_PARTICIPATING",
|
14 |
"106": "NON_PARTICIPATING",
|
15 |
"107": "NON_PARTICIPATING",
|
|
|
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",
|
|
|
38 |
"128": "NON_PARTICIPATING",
|
39 |
"129": "SUCCESS",
|
40 |
"13": "SUCCESS",
|
41 |
+
"130": "SUCCESS",
|
42 |
"131": "NON_PARTICIPATING",
|
43 |
"132": "NON_PARTICIPATING",
|
44 |
"133": "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": "NON_PARTICIPATING",
|
|
|
56 |
"144": "NON_PARTICIPATING",
|
57 |
"145": "NON_PARTICIPATING",
|
58 |
"146": "NON_PARTICIPATING",
|
59 |
+
"147": "SUCCESS",
|
60 |
+
"148": "SUCCESS",
|
61 |
"149": "NON_PARTICIPATING",
|
62 |
+
"15": "SUCCESS",
|
63 |
"150": "NON_PARTICIPATING",
|
64 |
"151": "NON_PARTICIPATING",
|
65 |
+
"152": "SUCCESS",
|
66 |
"153": "NON_PARTICIPATING",
|
67 |
"154": "NON_PARTICIPATING",
|
68 |
"155": "NON_PARTICIPATING",
|
69 |
+
"156": "NON_PARTICIPATING",
|
70 |
"157": "NON_PARTICIPATING",
|
71 |
"158": "NON_PARTICIPATING",
|
72 |
"159": "NON_PARTICIPATING",
|
73 |
+
"16": "SUCCESS",
|
74 |
"160": "NON_PARTICIPATING",
|
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": "NON_PARTICIPATING",
|
82 |
"168": "NON_PARTICIPATING",
|
83 |
"169": "NON_PARTICIPATING",
|
84 |
+
"17": "SUCCESS",
|
85 |
"170": "NON_PARTICIPATING",
|
86 |
"171": "NON_PARTICIPATING",
|
87 |
"172": "NON_PARTICIPATING",
|
88 |
+
"173": "SUCCESS",
|
89 |
"174": "NON_PARTICIPATING",
|
90 |
"175": "NON_PARTICIPATING",
|
91 |
"176": "NON_PARTICIPATING",
|
92 |
"177": "NON_PARTICIPATING",
|
93 |
"178": "NON_PARTICIPATING",
|
94 |
"179": "NON_PARTICIPATING",
|
95 |
+
"18": "SUCCESS",
|
96 |
"180": "NON_PARTICIPATING",
|
97 |
"181": "NON_PARTICIPATING",
|
98 |
"182": "NON_PARTICIPATING",
|
|
|
101 |
"185": "NON_PARTICIPATING",
|
102 |
"186": "NON_PARTICIPATING",
|
103 |
"187": "NON_PARTICIPATING",
|
104 |
+
"188": "SUCCESS",
|
105 |
"189": "NON_PARTICIPATING",
|
106 |
"19": "SUCCESS",
|
107 |
"190": "NON_PARTICIPATING",
|
|
|
110 |
"193": "NON_PARTICIPATING",
|
111 |
"194": "NON_PARTICIPATING",
|
112 |
"195": "NON_PARTICIPATING",
|
113 |
+
"196": "SUCCESS",
|
114 |
"197": "NON_PARTICIPATING",
|
115 |
"198": "NON_PARTICIPATING",
|
116 |
"199": "NON_PARTICIPATING",
|
117 |
+
"2": "SUCCESS",
|
118 |
"20": "SUCCESS",
|
119 |
"200": "NON_PARTICIPATING",
|
120 |
"201": "NON_PARTICIPATING",
|
121 |
+
"202": "SUCCESS",
|
122 |
"203": "NON_PARTICIPATING",
|
123 |
"204": "NON_PARTICIPATING",
|
124 |
"205": "NON_PARTICIPATING",
|
|
|
136 |
"216": "NON_PARTICIPATING",
|
137 |
"217": "SUCCESS",
|
138 |
"218": "NON_PARTICIPATING",
|
139 |
+
"219": "SUCCESS",
|
140 |
+
"22": "SUCCESS",
|
141 |
"220": "NON_PARTICIPATING",
|
142 |
"221": "NON_PARTICIPATING",
|
143 |
"222": "NON_PARTICIPATING",
|
|
|
155 |
"233": "NON_PARTICIPATING",
|
156 |
"234": "NON_PARTICIPATING",
|
157 |
"235": "NON_PARTICIPATING",
|
158 |
+
"236": "SUCCESS",
|
159 |
"237": "NON_PARTICIPATING",
|
160 |
"238": "NON_PARTICIPATING",
|
161 |
"239": "NON_PARTICIPATING",
|
162 |
+
"24": "SUCCESS",
|
163 |
"240": "NON_PARTICIPATING",
|
164 |
"241": "NON_PARTICIPATING",
|
165 |
"242": "NON_PARTICIPATING",
|
|
|
177 |
"253": "NON_PARTICIPATING",
|
178 |
"254": "NON_PARTICIPATING",
|
179 |
"255": "NON_PARTICIPATING",
|
180 |
+
"26": "SUCCESS",
|
181 |
"27": "SUCCESS",
|
182 |
"28": "SUCCESS",
|
183 |
+
"29": "SUCCESS",
|
184 |
"3": "NON_PARTICIPATING",
|
185 |
"30": "NON_PARTICIPATING",
|
186 |
+
"31": "SUCCESS",
|
187 |
"32": "NON_PARTICIPATING",
|
188 |
"33": "NON_PARTICIPATING",
|
189 |
"34": "SUCCESS",
|
190 |
+
"35": "SUCCESS",
|
191 |
"36": "NON_PARTICIPATING",
|
192 |
"37": "NON_PARTICIPATING",
|
193 |
"38": "SUCCESS",
|
194 |
+
"39": "NON_PARTICIPATING",
|
195 |
+
"4": "SUCCESS",
|
196 |
"40": "SUCCESS",
|
197 |
+
"41": "SUCCESS",
|
198 |
"42": "SUCCESS",
|
199 |
"43": "NON_PARTICIPATING",
|
200 |
"44": "SUCCESS",
|
201 |
+
"45": "SUCCESS",
|
202 |
+
"46": "SUCCESS",
|
203 |
"47": "NON_PARTICIPATING",
|
204 |
"48": "NON_PARTICIPATING",
|
205 |
+
"49": "SUCCESS",
|
206 |
"5": "SUCCESS",
|
207 |
"50": "NON_PARTICIPATING",
|
208 |
"51": "SUCCESS",
|
209 |
"52": "SUCCESS",
|
210 |
"53": "NON_PARTICIPATING",
|
211 |
+
"54": "SUCCESS",
|
212 |
+
"55": "SUCCESS",
|
213 |
+
"56": "SUCCESS",
|
214 |
"57": "NON_PARTICIPATING",
|
215 |
"58": "NON_PARTICIPATING",
|
216 |
"59": "SUCCESS",
|
217 |
"6": "NON_PARTICIPATING",
|
218 |
"60": "SUCCESS",
|
219 |
+
"61": "SUCCESS",
|
220 |
+
"62": "SUCCESS",
|
221 |
"63": "SUCCESS",
|
222 |
"64": "SUCCESS",
|
223 |
+
"65": "SUCCESS",
|
224 |
"66": "NON_PARTICIPATING",
|
225 |
+
"67": "SUCCESS",
|
226 |
"68": "NON_PARTICIPATING",
|
227 |
"69": "NON_PARTICIPATING",
|
228 |
+
"7": "SUCCESS",
|
229 |
"70": "NON_PARTICIPATING",
|
230 |
"71": "SUCCESS",
|
231 |
+
"72": "SUCCESS",
|
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": "SUCCESS",
|
240 |
+
"80": "SUCCESS",
|
241 |
"81": "NON_PARTICIPATING",
|
242 |
"82": "NON_PARTICIPATING",
|
243 |
+
"83": "SUCCESS",
|
244 |
+
"84": "SUCCESS",
|
245 |
"85": "SUCCESS",
|
246 |
"86": "NON_PARTICIPATING",
|
247 |
+
"87": "SUCCESS",
|
248 |
+
"88": "SUCCESS",
|
249 |
+
"89": "SUCCESS",
|
250 |
"9": "SUCCESS",
|
251 |
+
"90": "SUCCESS",
|
252 |
"91": "SUCCESS",
|
253 |
"92": "NON_PARTICIPATING",
|
254 |
"93": "NON_PARTICIPATING",
|
255 |
+
"94": "SUCCESS",
|
256 |
"95": "SUCCESS",
|
257 |
"96": "NON_PARTICIPATING",
|
258 |
+
"97": "SUCCESS",
|
259 |
+
"98": "SUCCESS",
|
260 |
"99": "NON_PARTICIPATING"
|
261 |
},
|
262 |
"architectures": [
|
|
|
275 |
"initializer_range": 0.02,
|
276 |
"inner_step": 0,
|
277 |
"inner_steps": 0,
|
278 |
+
"last_allreduce_block": 5673692,
|
279 |
"layer_norm_epsilon": 1e-05,
|
280 |
"model_type": "gpt_optimized",
|
281 |
"n_embd": 1280,
|
inner_optimizer.pt
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 8081782026
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:533d1d831ec047a0ed40be45d4a2e627a0059bf63ea273b28f24b7f2c7a557af
|
3 |
size 8081782026
|
model.safetensors
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 4040701744
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:568529a818118b1adfcffb9e7bac688d0564b1d8763d605cad7a5fc413fc5f40
|
3 |
size 4040701744
|
outer_optimizer.pt
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 4040805354
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:42b9752f78b6fc32d53d77af4bf8217689b01a94c64bc1557e07ee5a0de21a68
|
3 |
size 4040805354
|