Run 4. Outer Step 99. Inner Step 0. Peers 67.
Browse files- config.json +22 -22
- inner_optimizer.pt +1 -1
- model.safetensors +1 -1
- outer_optimizer.pt +1 -1
config.json
CHANGED
@@ -6,10 +6,10 @@
|
|
6 |
"1": "SUCCESS",
|
7 |
"10": "NON_PARTICIPATING",
|
8 |
"100": "NON_PARTICIPATING",
|
9 |
-
"101": "
|
10 |
"102": "NON_PARTICIPATING",
|
11 |
"103": "NON_PARTICIPATING",
|
12 |
-
"104": "
|
13 |
"105": "NON_PARTICIPATING",
|
14 |
"106": "NON_PARTICIPATING",
|
15 |
"107": "NON_PARTICIPATING",
|
@@ -36,17 +36,17 @@
|
|
36 |
"126": "NON_PARTICIPATING",
|
37 |
"127": "NON_PARTICIPATING",
|
38 |
"128": "NON_PARTICIPATING",
|
39 |
-
"129": "
|
40 |
"13": "SUCCESS",
|
41 |
"130": "NON_PARTICIPATING",
|
42 |
"131": "NON_PARTICIPATING",
|
43 |
-
"132": "
|
44 |
"133": "NON_PARTICIPATING",
|
45 |
"134": "NON_PARTICIPATING",
|
46 |
"135": "NON_PARTICIPATING",
|
47 |
"136": "NON_PARTICIPATING",
|
48 |
"137": "NON_PARTICIPATING",
|
49 |
-
"138": "
|
50 |
"139": "NON_PARTICIPATING",
|
51 |
"14": "SUCCESS",
|
52 |
"140": "NON_PARTICIPATING",
|
@@ -65,7 +65,7 @@
|
|
65 |
"152": "NON_PARTICIPATING",
|
66 |
"153": "NON_PARTICIPATING",
|
67 |
"154": "NON_PARTICIPATING",
|
68 |
-
"155": "
|
69 |
"156": "NON_PARTICIPATING",
|
70 |
"157": "NON_PARTICIPATING",
|
71 |
"158": "NON_PARTICIPATING",
|
@@ -78,20 +78,20 @@
|
|
78 |
"164": "NON_PARTICIPATING",
|
79 |
"165": "NON_PARTICIPATING",
|
80 |
"166": "NON_PARTICIPATING",
|
81 |
-
"167": "
|
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": "
|
89 |
"174": "NON_PARTICIPATING",
|
90 |
"175": "NON_PARTICIPATING",
|
91 |
"176": "NON_PARTICIPATING",
|
92 |
"177": "NON_PARTICIPATING",
|
93 |
"178": "NON_PARTICIPATING",
|
94 |
-
"179": "
|
95 |
"18": "SUCCESS",
|
96 |
"180": "NON_PARTICIPATING",
|
97 |
"181": "NON_PARTICIPATING",
|
@@ -106,22 +106,22 @@
|
|
106 |
"19": "NON_PARTICIPATING",
|
107 |
"190": "NON_PARTICIPATING",
|
108 |
"191": "NON_PARTICIPATING",
|
109 |
-
"192": "
|
110 |
"193": "NON_PARTICIPATING",
|
111 |
"194": "NON_PARTICIPATING",
|
112 |
"195": "NON_PARTICIPATING",
|
113 |
"196": "NON_PARTICIPATING",
|
114 |
-
"197": "
|
115 |
"198": "NON_PARTICIPATING",
|
116 |
"199": "NON_PARTICIPATING",
|
117 |
"2": "SUCCESS",
|
118 |
-
"20": "
|
119 |
-
"200": "
|
120 |
"201": "NON_PARTICIPATING",
|
121 |
-
"202": "
|
122 |
"203": "NON_PARTICIPATING",
|
123 |
"204": "NON_PARTICIPATING",
|
124 |
-
"205": "
|
125 |
"206": "NON_PARTICIPATING",
|
126 |
"207": "NON_PARTICIPATING",
|
127 |
"208": "NON_PARTICIPATING",
|
@@ -151,18 +151,18 @@
|
|
151 |
"23": "SUCCESS",
|
152 |
"230": "NON_PARTICIPATING",
|
153 |
"231": "NON_PARTICIPATING",
|
154 |
-
"232": "
|
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": "SUCCESS",
|
163 |
"240": "NON_PARTICIPATING",
|
164 |
"241": "NON_PARTICIPATING",
|
165 |
-
"242": "
|
166 |
"243": "NON_PARTICIPATING",
|
167 |
"244": "NON_PARTICIPATING",
|
168 |
"245": "NON_PARTICIPATING",
|
@@ -202,7 +202,7 @@
|
|
202 |
"46": "SUCCESS",
|
203 |
"47": "NON_PARTICIPATING",
|
204 |
"48": "NON_PARTICIPATING",
|
205 |
-
"49": "
|
206 |
"5": "NON_PARTICIPATING",
|
207 |
"50": "SUCCESS",
|
208 |
"51": "SUCCESS",
|
@@ -227,7 +227,7 @@
|
|
227 |
"69": "SUCCESS",
|
228 |
"7": "SUCCESS",
|
229 |
"70": "NON_PARTICIPATING",
|
230 |
-
"71": "
|
231 |
"72": "NON_PARTICIPATING",
|
232 |
"73": "SUCCESS",
|
233 |
"74": "NON_PARTICIPATING",
|
@@ -253,7 +253,7 @@
|
|
253 |
"92": "NON_PARTICIPATING",
|
254 |
"93": "SUCCESS",
|
255 |
"94": "SUCCESS",
|
256 |
-
"95": "
|
257 |
"96": "NON_PARTICIPATING",
|
258 |
"97": "NON_PARTICIPATING",
|
259 |
"98": "SUCCESS",
|
@@ -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,
|
|
|
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",
|
|
|
36 |
"126": "NON_PARTICIPATING",
|
37 |
"127": "NON_PARTICIPATING",
|
38 |
"128": "NON_PARTICIPATING",
|
39 |
+
"129": "SUCCESS",
|
40 |
"13": "SUCCESS",
|
41 |
"130": "NON_PARTICIPATING",
|
42 |
"131": "NON_PARTICIPATING",
|
43 |
+
"132": "SUCCESS",
|
44 |
"133": "NON_PARTICIPATING",
|
45 |
"134": "NON_PARTICIPATING",
|
46 |
"135": "NON_PARTICIPATING",
|
47 |
"136": "NON_PARTICIPATING",
|
48 |
"137": "NON_PARTICIPATING",
|
49 |
+
"138": "NON_PARTICIPATING",
|
50 |
"139": "NON_PARTICIPATING",
|
51 |
"14": "SUCCESS",
|
52 |
"140": "NON_PARTICIPATING",
|
|
|
65 |
"152": "NON_PARTICIPATING",
|
66 |
"153": "NON_PARTICIPATING",
|
67 |
"154": "NON_PARTICIPATING",
|
68 |
+
"155": "SUCCESS",
|
69 |
"156": "NON_PARTICIPATING",
|
70 |
"157": "NON_PARTICIPATING",
|
71 |
"158": "NON_PARTICIPATING",
|
|
|
78 |
"164": "NON_PARTICIPATING",
|
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": "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",
|
95 |
"18": "SUCCESS",
|
96 |
"180": "NON_PARTICIPATING",
|
97 |
"181": "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": "NON_PARTICIPATING",
|
113 |
"196": "NON_PARTICIPATING",
|
114 |
+
"197": "SUCCESS",
|
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",
|
125 |
"206": "NON_PARTICIPATING",
|
126 |
"207": "NON_PARTICIPATING",
|
127 |
"208": "NON_PARTICIPATING",
|
|
|
151 |
"23": "SUCCESS",
|
152 |
"230": "NON_PARTICIPATING",
|
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": "NON_PARTICIPATING",
|
161 |
"239": "NON_PARTICIPATING",
|
162 |
"24": "SUCCESS",
|
163 |
"240": "NON_PARTICIPATING",
|
164 |
"241": "NON_PARTICIPATING",
|
165 |
+
"242": "SUCCESS",
|
166 |
"243": "NON_PARTICIPATING",
|
167 |
"244": "NON_PARTICIPATING",
|
168 |
"245": "NON_PARTICIPATING",
|
|
|
202 |
"46": "SUCCESS",
|
203 |
"47": "NON_PARTICIPATING",
|
204 |
"48": "NON_PARTICIPATING",
|
205 |
+
"49": "SUCCESS",
|
206 |
"5": "NON_PARTICIPATING",
|
207 |
"50": "SUCCESS",
|
208 |
"51": "SUCCESS",
|
|
|
227 |
"69": "SUCCESS",
|
228 |
"7": "SUCCESS",
|
229 |
"70": "NON_PARTICIPATING",
|
230 |
+
"71": "SUCCESS",
|
231 |
"72": "NON_PARTICIPATING",
|
232 |
"73": "SUCCESS",
|
233 |
"74": "NON_PARTICIPATING",
|
|
|
253 |
"92": "NON_PARTICIPATING",
|
254 |
"93": "SUCCESS",
|
255 |
"94": "SUCCESS",
|
256 |
+
"95": "SUCCESS",
|
257 |
"96": "NON_PARTICIPATING",
|
258 |
"97": "NON_PARTICIPATING",
|
259 |
"98": "SUCCESS",
|
|
|
275 |
"initializer_range": 0.02,
|
276 |
"inner_step": 0,
|
277 |
"inner_steps": 0,
|
278 |
+
"last_allreduce_block": 5694192,
|
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 sha256:4988587df8772e23c704f6197e0ec708bac4f5821f6c36d05ed7aaaec8c8caba
|
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 sha256:680342a859ed8575c08833f1fb3326f90bb3cc8ffa66f1a4fa28cc18410ca358
|
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:c42051bbb9023a248aa900f19209b9c68299d355c698ad4191d6c177676d5b87
|
3 |
size 4040805354
|