Run 4. Outer Step 2. Inner Step 0. Peers 61.
Browse files- config.json +37 -37
- inner_optimizer.pt +1 -1
- model.safetensors +1 -1
- outer_optimizer.pt +1 -1
config.json
CHANGED
@@ -1,21 +1,21 @@
|
|
1 |
{
|
2 |
-
"_name_or_path": "
|
3 |
"activation_function": "gelu_new",
|
4 |
"all_reduce_scores": {
|
5 |
"0": "NON_PARTICIPATING",
|
6 |
-
"1": "
|
7 |
-
"10": "
|
8 |
"100": "NON_PARTICIPATING",
|
9 |
-
"101": "
|
10 |
"102": "NON_PARTICIPATING",
|
11 |
"103": "NON_PARTICIPATING",
|
12 |
"104": "NON_PARTICIPATING",
|
13 |
"105": "NON_PARTICIPATING",
|
14 |
"106": "NON_PARTICIPATING",
|
15 |
-
"107": "
|
16 |
"108": "NON_PARTICIPATING",
|
17 |
-
"109": "
|
18 |
-
"11": "
|
19 |
"110": "SUCCESS",
|
20 |
"111": "NON_PARTICIPATING",
|
21 |
"112": "NON_PARTICIPATING",
|
@@ -27,9 +27,9 @@
|
|
27 |
"118": "NON_PARTICIPATING",
|
28 |
"119": "NON_PARTICIPATING",
|
29 |
"12": "NON_PARTICIPATING",
|
30 |
-
"120": "
|
31 |
"121": "NON_PARTICIPATING",
|
32 |
-
"122": "
|
33 |
"123": "NON_PARTICIPATING",
|
34 |
"124": "NON_PARTICIPATING",
|
35 |
"125": "SUCCESS",
|
@@ -51,13 +51,13 @@
|
|
51 |
"14": "NON_PARTICIPATING",
|
52 |
"140": "NON_PARTICIPATING",
|
53 |
"141": "NON_PARTICIPATING",
|
54 |
-
"142": "
|
55 |
"143": "NON_PARTICIPATING",
|
56 |
"144": "NON_PARTICIPATING",
|
57 |
"145": "NON_PARTICIPATING",
|
58 |
"146": "NON_PARTICIPATING",
|
59 |
"147": "SUCCESS",
|
60 |
-
"148": "
|
61 |
"149": "NON_PARTICIPATING",
|
62 |
"15": "NON_PARTICIPATING",
|
63 |
"150": "NON_PARTICIPATING",
|
@@ -89,8 +89,8 @@
|
|
89 |
"174": "NON_PARTICIPATING",
|
90 |
"175": "NON_PARTICIPATING",
|
91 |
"176": "NON_PARTICIPATING",
|
92 |
-
"177": "
|
93 |
-
"178": "
|
94 |
"179": "NON_PARTICIPATING",
|
95 |
"18": "NON_PARTICIPATING",
|
96 |
"180": "NON_PARTICIPATING",
|
@@ -106,16 +106,16 @@
|
|
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",
|
116 |
"199": "NON_PARTICIPATING",
|
117 |
"2": "SUCCESS",
|
118 |
-
"20": "
|
119 |
"200": "NON_PARTICIPATING",
|
120 |
"201": "NON_PARTICIPATING",
|
121 |
"202": "NON_PARTICIPATING",
|
@@ -131,24 +131,24 @@
|
|
131 |
"211": "NON_PARTICIPATING",
|
132 |
"212": "NON_PARTICIPATING",
|
133 |
"213": "NON_PARTICIPATING",
|
134 |
-
"214": "
|
135 |
"215": "NON_PARTICIPATING",
|
136 |
-
"216": "
|
137 |
"217": "NON_PARTICIPATING",
|
138 |
-
"218": "
|
139 |
"219": "NON_PARTICIPATING",
|
140 |
"22": "SUCCESS",
|
141 |
"220": "NON_PARTICIPATING",
|
142 |
"221": "NON_PARTICIPATING",
|
143 |
"222": "SUCCESS",
|
144 |
-
"223": "
|
145 |
"224": "NON_PARTICIPATING",
|
146 |
-
"225": "
|
147 |
"226": "NON_PARTICIPATING",
|
148 |
"227": "NON_PARTICIPATING",
|
149 |
"228": "NON_PARTICIPATING",
|
150 |
"229": "NON_PARTICIPATING",
|
151 |
-
"23": "
|
152 |
"230": "NON_PARTICIPATING",
|
153 |
"231": "NON_PARTICIPATING",
|
154 |
"232": "NON_PARTICIPATING",
|
@@ -176,8 +176,8 @@
|
|
176 |
"252": "NON_PARTICIPATING",
|
177 |
"253": "NON_PARTICIPATING",
|
178 |
"254": "NON_PARTICIPATING",
|
179 |
-
"255": "
|
180 |
-
"26": "
|
181 |
"27": "SUCCESS",
|
182 |
"28": "NON_PARTICIPATING",
|
183 |
"29": "NON_PARTICIPATING",
|
@@ -189,9 +189,9 @@
|
|
189 |
"34": "NON_PARTICIPATING",
|
190 |
"35": "NON_PARTICIPATING",
|
191 |
"36": "NON_PARTICIPATING",
|
192 |
-
"37": "
|
193 |
"38": "NON_PARTICIPATING",
|
194 |
-
"39": "
|
195 |
"4": "NON_PARTICIPATING",
|
196 |
"40": "NON_PARTICIPATING",
|
197 |
"41": "SUCCESS",
|
@@ -202,27 +202,27 @@
|
|
202 |
"46": "NON_PARTICIPATING",
|
203 |
"47": "NON_PARTICIPATING",
|
204 |
"48": "NON_PARTICIPATING",
|
205 |
-
"49": "
|
206 |
"5": "NON_PARTICIPATING",
|
207 |
-
"50": "
|
208 |
"51": "SUCCESS",
|
209 |
"52": "NON_PARTICIPATING",
|
210 |
-
"53": "
|
211 |
"54": "NON_PARTICIPATING",
|
212 |
"55": "SUCCESS",
|
213 |
-
"56": "
|
214 |
"57": "NON_PARTICIPATING",
|
215 |
"58": "SUCCESS",
|
216 |
"59": "NON_PARTICIPATING",
|
217 |
"6": "NON_PARTICIPATING",
|
218 |
-
"60": "
|
219 |
"61": "NON_PARTICIPATING",
|
220 |
-
"62": "
|
221 |
-
"63": "
|
222 |
"64": "NON_PARTICIPATING",
|
223 |
"65": "NON_PARTICIPATING",
|
224 |
"66": "NON_PARTICIPATING",
|
225 |
-
"67": "
|
226 |
"68": "NON_PARTICIPATING",
|
227 |
"69": "NON_PARTICIPATING",
|
228 |
"7": "NON_PARTICIPATING",
|
@@ -234,7 +234,7 @@
|
|
234 |
"75": "SUCCESS",
|
235 |
"76": "NON_PARTICIPATING",
|
236 |
"77": "NON_PARTICIPATING",
|
237 |
-
"78": "
|
238 |
"79": "NON_PARTICIPATING",
|
239 |
"8": "NON_PARTICIPATING",
|
240 |
"80": "SUCCESS",
|
@@ -257,7 +257,7 @@
|
|
257 |
"96": "NON_PARTICIPATING",
|
258 |
"97": "NON_PARTICIPATING",
|
259 |
"98": "NON_PARTICIPATING",
|
260 |
-
"99": "
|
261 |
},
|
262 |
"architectures": [
|
263 |
"GPTOptim"
|
@@ -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": "wandb01/c66-h23",
|
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",
|
11 |
"103": "NON_PARTICIPATING",
|
12 |
"104": "NON_PARTICIPATING",
|
13 |
"105": "NON_PARTICIPATING",
|
14 |
"106": "NON_PARTICIPATING",
|
15 |
+
"107": "SUCCESS",
|
16 |
"108": "NON_PARTICIPATING",
|
17 |
+
"109": "SUCCESS",
|
18 |
+
"11": "SUCCESS",
|
19 |
"110": "SUCCESS",
|
20 |
"111": "NON_PARTICIPATING",
|
21 |
"112": "NON_PARTICIPATING",
|
|
|
27 |
"118": "NON_PARTICIPATING",
|
28 |
"119": "NON_PARTICIPATING",
|
29 |
"12": "NON_PARTICIPATING",
|
30 |
+
"120": "NON_PARTICIPATING",
|
31 |
"121": "NON_PARTICIPATING",
|
32 |
+
"122": "SUCCESS",
|
33 |
"123": "NON_PARTICIPATING",
|
34 |
"124": "NON_PARTICIPATING",
|
35 |
"125": "SUCCESS",
|
|
|
51 |
"14": "NON_PARTICIPATING",
|
52 |
"140": "NON_PARTICIPATING",
|
53 |
"141": "NON_PARTICIPATING",
|
54 |
+
"142": "SUCCESS",
|
55 |
"143": "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": "NON_PARTICIPATING",
|
63 |
"150": "NON_PARTICIPATING",
|
|
|
89 |
"174": "NON_PARTICIPATING",
|
90 |
"175": "NON_PARTICIPATING",
|
91 |
"176": "NON_PARTICIPATING",
|
92 |
+
"177": "SUCCESS",
|
93 |
+
"178": "SUCCESS",
|
94 |
"179": "NON_PARTICIPATING",
|
95 |
"18": "NON_PARTICIPATING",
|
96 |
"180": "NON_PARTICIPATING",
|
|
|
106 |
"19": "NON_PARTICIPATING",
|
107 |
"190": "NON_PARTICIPATING",
|
108 |
"191": "NON_PARTICIPATING",
|
109 |
+
"192": "SUCCESS",
|
110 |
"193": "NON_PARTICIPATING",
|
111 |
"194": "NON_PARTICIPATING",
|
112 |
+
"195": "SUCCESS",
|
113 |
"196": "NON_PARTICIPATING",
|
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": "NON_PARTICIPATING",
|
|
|
131 |
"211": "NON_PARTICIPATING",
|
132 |
"212": "NON_PARTICIPATING",
|
133 |
"213": "NON_PARTICIPATING",
|
134 |
+
"214": "SUCCESS",
|
135 |
"215": "NON_PARTICIPATING",
|
136 |
+
"216": "NON_PARTICIPATING",
|
137 |
"217": "NON_PARTICIPATING",
|
138 |
+
"218": "SUCCESS",
|
139 |
"219": "NON_PARTICIPATING",
|
140 |
"22": "SUCCESS",
|
141 |
"220": "NON_PARTICIPATING",
|
142 |
"221": "NON_PARTICIPATING",
|
143 |
"222": "SUCCESS",
|
144 |
+
"223": "SUCCESS",
|
145 |
"224": "NON_PARTICIPATING",
|
146 |
+
"225": "SUCCESS",
|
147 |
"226": "NON_PARTICIPATING",
|
148 |
"227": "NON_PARTICIPATING",
|
149 |
"228": "NON_PARTICIPATING",
|
150 |
"229": "NON_PARTICIPATING",
|
151 |
+
"23": "SUCCESS",
|
152 |
"230": "NON_PARTICIPATING",
|
153 |
"231": "NON_PARTICIPATING",
|
154 |
"232": "NON_PARTICIPATING",
|
|
|
176 |
"252": "NON_PARTICIPATING",
|
177 |
"253": "NON_PARTICIPATING",
|
178 |
"254": "NON_PARTICIPATING",
|
179 |
+
"255": "SUCCESS",
|
180 |
+
"26": "SUCCESS",
|
181 |
"27": "SUCCESS",
|
182 |
"28": "NON_PARTICIPATING",
|
183 |
"29": "NON_PARTICIPATING",
|
|
|
189 |
"34": "NON_PARTICIPATING",
|
190 |
"35": "NON_PARTICIPATING",
|
191 |
"36": "NON_PARTICIPATING",
|
192 |
+
"37": "SUCCESS",
|
193 |
"38": "NON_PARTICIPATING",
|
194 |
+
"39": "SUCCESS",
|
195 |
"4": "NON_PARTICIPATING",
|
196 |
"40": "NON_PARTICIPATING",
|
197 |
"41": "SUCCESS",
|
|
|
202 |
"46": "NON_PARTICIPATING",
|
203 |
"47": "NON_PARTICIPATING",
|
204 |
"48": "NON_PARTICIPATING",
|
205 |
+
"49": "SUCCESS",
|
206 |
"5": "NON_PARTICIPATING",
|
207 |
+
"50": "SUCCESS",
|
208 |
"51": "SUCCESS",
|
209 |
"52": "NON_PARTICIPATING",
|
210 |
+
"53": "SUCCESS",
|
211 |
"54": "NON_PARTICIPATING",
|
212 |
"55": "SUCCESS",
|
213 |
+
"56": "FAIL",
|
214 |
"57": "NON_PARTICIPATING",
|
215 |
"58": "SUCCESS",
|
216 |
"59": "NON_PARTICIPATING",
|
217 |
"6": "NON_PARTICIPATING",
|
218 |
+
"60": "SUCCESS",
|
219 |
"61": "NON_PARTICIPATING",
|
220 |
+
"62": "NON_PARTICIPATING",
|
221 |
+
"63": "SUCCESS",
|
222 |
"64": "NON_PARTICIPATING",
|
223 |
"65": "NON_PARTICIPATING",
|
224 |
"66": "NON_PARTICIPATING",
|
225 |
+
"67": "SUCCESS",
|
226 |
"68": "NON_PARTICIPATING",
|
227 |
"69": "NON_PARTICIPATING",
|
228 |
"7": "NON_PARTICIPATING",
|
|
|
234 |
"75": "SUCCESS",
|
235 |
"76": "NON_PARTICIPATING",
|
236 |
"77": "NON_PARTICIPATING",
|
237 |
+
"78": "SUCCESS",
|
238 |
"79": "NON_PARTICIPATING",
|
239 |
"8": "NON_PARTICIPATING",
|
240 |
"80": "SUCCESS",
|
|
|
257 |
"96": "NON_PARTICIPATING",
|
258 |
"97": "NON_PARTICIPATING",
|
259 |
"98": "NON_PARTICIPATING",
|
260 |
+
"99": "FAIL"
|
261 |
},
|
262 |
"architectures": [
|
263 |
"GPTOptim"
|
|
|
275 |
"initializer_range": 0.02,
|
276 |
"inner_step": 0,
|
277 |
"inner_steps": 0,
|
278 |
+
"last_allreduce_block": 5477809,
|
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 sha256:62f60901918a8f5efcc10e97b1b0169307c8bcc3e46c0511abb46e11ffcb8ceb
|
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 sha256:68389cb17e9d9ca041ff1881731d614e9a6a76858dd0ff364aafaafe78d601eb
|
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 sha256:3a806d37298938565e45c6fc97b7bd31f616b43960f7e6fab76e7e7448f5cc9b
|
3 |
size 4040805354
|