Epoch 1675. Batch Size 58416. Peers 100.

config.json

@@ -3,26 +3,26 @@
   "activation_function": "gelu_new",
   "all_reduce_scores": {
     "0": "NON_PARTICIPATING",
-    "1": "NON_PARTICIPATING",
-    "10": "NON_PARTICIPATING",
+    "1": "SUCCESS",
+    "10": "SUCCESS",
     "100": "NON_PARTICIPATING",
-    "101": "NON_PARTICIPATING",
+    "101": "SUCCESS",
     "102": "NON_PARTICIPATING",
     "103": "NON_PARTICIPATING",
     "104": "NON_PARTICIPATING",
-    "105": "NON_PARTICIPATING",
+    "105": "SUCCESS",
     "106": "SUCCESS",
     "107": "SUCCESS",
     "108": "NON_PARTICIPATING",
-    "109": "NON_PARTICIPATING",
-    "11": "SUCCESS",
+    "109": "SUCCESS",
+    "11": "NON_PARTICIPATING",
     "110": "NON_PARTICIPATING",
-    "111": "SUCCESS",
+    "111": "NON_PARTICIPATING",
     "112": "NON_PARTICIPATING",
-    "113": "SUCCESS",
+    "113": "NON_PARTICIPATING",
     "114": "NON_PARTICIPATING",
-    "115": "NON_PARTICIPATING",
-    "116": "NON_PARTICIPATING",
+    "115": "SUCCESS",
+    "116": "SUCCESS",
     "117": "NON_PARTICIPATING",
     "118": "SUCCESS",
     "119": "NON_PARTICIPATING",
@@ -30,14 +30,14 @@
     "120": "NON_PARTICIPATING",
     "121": "SUCCESS",
     "122": "SUCCESS",
-    "123": "SUCCESS",
+    "123": "NON_PARTICIPATING",
     "124": "NON_PARTICIPATING",
     "125": "NON_PARTICIPATING",
     "126": "SUCCESS",
     "127": "NON_PARTICIPATING",
-    "128": "NON_PARTICIPATING",
+    "128": "SUCCESS",
     "129": "NON_PARTICIPATING",
-    "13": "NON_PARTICIPATING",
+    "13": "SUCCESS",
     "130": "NON_PARTICIPATING",
     "131": "NON_PARTICIPATING",
     "132": "SUCCESS",
@@ -45,99 +45,99 @@
     "134": "SUCCESS",
     "135": "NON_PARTICIPATING",
     "136": "NON_PARTICIPATING",
-    "137": "NON_PARTICIPATING",
-    "138": "SUCCESS",
-    "139": "SUCCESS",
-    "14": "NON_PARTICIPATING",
+    "137": "SUCCESS",
+    "138": "NON_PARTICIPATING",
+    "139": "NON_PARTICIPATING",
+    "14": "SUCCESS",
     "140": "SUCCESS",
     "141": "NON_PARTICIPATING",
     "142": "NON_PARTICIPATING",
     "143": "NON_PARTICIPATING",
     "144": "NON_PARTICIPATING",
-    "145": "NON_PARTICIPATING",
+    "145": "SUCCESS",
     "146": "NON_PARTICIPATING",
     "147": "NON_PARTICIPATING",
     "148": "NON_PARTICIPATING",
     "149": "SUCCESS",
-    "15": "NON_PARTICIPATING",
+    "15": "SUCCESS",
     "150": "SUCCESS",
     "151": "NON_PARTICIPATING",
-    "152": "SUCCESS",
+    "152": "NON_PARTICIPATING",
     "153": "NON_PARTICIPATING",
     "154": "NON_PARTICIPATING",
     "155": "NON_PARTICIPATING",
-    "156": "SUCCESS",
-    "157": "NON_PARTICIPATING",
+    "156": "NON_PARTICIPATING",
+    "157": "SUCCESS",
     "158": "SUCCESS",
-    "159": "SUCCESS",
-    "16": "SUCCESS",
+    "159": "NON_PARTICIPATING",
+    "16": "NON_PARTICIPATING",
     "160": "NON_PARTICIPATING",
     "161": "SUCCESS",
     "162": "SUCCESS",
-    "163": "NON_PARTICIPATING",
-    "164": "NON_PARTICIPATING",
+    "163": "SUCCESS",
+    "164": "SUCCESS",
     "165": "NON_PARTICIPATING",
     "166": "SUCCESS",
-    "167": "NON_PARTICIPATING",
+    "167": "SUCCESS",
     "168": "NON_PARTICIPATING",
     "169": "NON_PARTICIPATING",
-    "17": "SUCCESS",
+    "17": "NON_PARTICIPATING",
     "170": "NON_PARTICIPATING",
     "171": "SUCCESS",
     "172": "NON_PARTICIPATING",
     "173": "SUCCESS",
     "174": "SUCCESS",
     "175": "SUCCESS",
-    "176": "SUCCESS",
+    "176": "NON_PARTICIPATING",
     "177": "NON_PARTICIPATING",
     "178": "NON_PARTICIPATING",
     "179": "NON_PARTICIPATING",
     "18": "SUCCESS",
-    "180": "SUCCESS",
-    "181": "SUCCESS",
-    "182": "SUCCESS",
+    "180": "NON_PARTICIPATING",
+    "181": "NON_PARTICIPATING",
+    "182": "NON_PARTICIPATING",
     "183": "NON_PARTICIPATING",
     "184": "SUCCESS",
     "185": "SUCCESS",
-    "186": "NON_PARTICIPATING",
+    "186": "SUCCESS",
     "187": "NON_PARTICIPATING",
-    "188": "NON_PARTICIPATING",
-    "189": "SUCCESS",
-    "19": "NON_PARTICIPATING",
-    "190": "NON_PARTICIPATING",
-    "191": "NON_PARTICIPATING",
+    "188": "SUCCESS",
+    "189": "NON_PARTICIPATING",
+    "19": "SUCCESS",
+    "190": "SUCCESS",
+    "191": "SUCCESS",
     "192": "NON_PARTICIPATING",
     "193": "NON_PARTICIPATING",
-    "194": "SUCCESS",
+    "194": "NON_PARTICIPATING",
     "195": "SUCCESS",
-    "196": "SUCCESS",
+    "196": "NON_PARTICIPATING",
     "197": "NON_PARTICIPATING",
     "198": "SUCCESS",
     "199": "SUCCESS",
     "2": "NON_PARTICIPATING",
     "20": "SUCCESS",
-    "200": "SUCCESS",
+    "200": "NON_PARTICIPATING",
     "201": "SUCCESS",
     "202": "NON_PARTICIPATING",
-    "203": "SUCCESS",
+    "203": "NON_PARTICIPATING",
     "204": "NON_PARTICIPATING",
     "205": "SUCCESS",
-    "206": "NON_PARTICIPATING",
+    "206": "SUCCESS",
     "207": "NON_PARTICIPATING",
-    "208": "SUCCESS",
-    "209": "NON_PARTICIPATING",
+    "208": "NON_PARTICIPATING",
+    "209": "SUCCESS",
     "21": "NON_PARTICIPATING",
     "210": "NON_PARTICIPATING",
     "211": "SUCCESS",
-    "212": "NON_PARTICIPATING",
-    "213": "SUCCESS",
+    "212": "SUCCESS",
+    "213": "NON_PARTICIPATING",
     "214": "NON_PARTICIPATING",
-    "215": "SUCCESS",
-    "216": "NON_PARTICIPATING",
+    "215": "NON_PARTICIPATING",
+    "216": "SUCCESS",
     "217": "SUCCESS",
-    "218": "NON_PARTICIPATING",
+    "218": "SUCCESS",
     "219": "SUCCESS",
-    "22": "SUCCESS",
+    "22": "NON_PARTICIPATING",
     "220": "SUCCESS",
     "221": "NON_PARTICIPATING",
     "222": "NON_PARTICIPATING",
@@ -151,36 +151,36 @@
     "23": "NON_PARTICIPATING",
     "230": "NON_PARTICIPATING",
     "231": "SUCCESS",
-    "232": "NON_PARTICIPATING",
+    "232": "SUCCESS",
     "233": "SUCCESS",
     "234": "SUCCESS",
-    "235": "NON_PARTICIPATING",
-    "236": "SUCCESS",
-    "237": "SUCCESS",
-    "238": "SUCCESS",
-    "239": "NON_PARTICIPATING",
+    "235": "SUCCESS",
+    "236": "NON_PARTICIPATING",
+    "237": "NON_PARTICIPATING",
+    "238": "NON_PARTICIPATING",
+    "239": "SUCCESS",
     "24": "SUCCESS",
     "240": "NON_PARTICIPATING",
     "241": "SUCCESS",
     "242": "NON_PARTICIPATING",
-    "243": "NON_PARTICIPATING",
-    "244": "NON_PARTICIPATING",
+    "243": "SUCCESS",
+    "244": "SUCCESS",
     "245": "SUCCESS",
-    "246": "SUCCESS",
-    "247": "SUCCESS",
-    "248": "SUCCESS",
-    "249": "SUCCESS",
+    "246": "NON_PARTICIPATING",
+    "247": "NON_PARTICIPATING",
+    "248": "NON_PARTICIPATING",
+    "249": "NON_PARTICIPATING",
     "25": "SUCCESS",
     "250": "NON_PARTICIPATING",
     "251": "NON_PARTICIPATING",
     "252": "NON_PARTICIPATING",
     "253": "NON_PARTICIPATING",
     "254": "NON_PARTICIPATING",
-    "255": "NON_PARTICIPATING",
-    "26": "NON_PARTICIPATING",
+    "255": "SUCCESS",
+    "26": "SUCCESS",
     "27": "NON_PARTICIPATING",
-    "28": "NON_PARTICIPATING",
-    "29": "SUCCESS",
+    "28": "SUCCESS",
+    "29": "NON_PARTICIPATING",
     "3": "NON_PARTICIPATING",
     "30": "NON_PARTICIPATING",
     "31": "SUCCESS",
@@ -188,12 +188,12 @@
     "33": "NON_PARTICIPATING",
     "34": "NON_PARTICIPATING",
     "35": "NON_PARTICIPATING",
-    "36": "SUCCESS",
-    "37": "NON_PARTICIPATING",
+    "36": "NON_PARTICIPATING",
+    "37": "SUCCESS",
     "38": "NON_PARTICIPATING",
-    "39": "SUCCESS",
+    "39": "NON_PARTICIPATING",
     "4": "SUCCESS",
-    "40": "SUCCESS",
+    "40": "NON_PARTICIPATING",
     "41": "NON_PARTICIPATING",
     "42": "NON_PARTICIPATING",
     "43": "NON_PARTICIPATING",
@@ -202,62 +202,62 @@
     "46": "SUCCESS",
     "47": "NON_PARTICIPATING",
     "48": "NON_PARTICIPATING",
-    "49": "NON_PARTICIPATING",
+    "49": "SUCCESS",
     "5": "NON_PARTICIPATING",
     "50": "NON_PARTICIPATING",
-    "51": "NON_PARTICIPATING",
+    "51": "SUCCESS",
     "52": "NON_PARTICIPATING",
     "53": "SUCCESS",
     "54": "NON_PARTICIPATING",
     "55": "NON_PARTICIPATING",
-    "56": "SUCCESS",
-    "57": "SUCCESS",
+    "56": "NON_PARTICIPATING",
+    "57": "NON_PARTICIPATING",
     "58": "SUCCESS",
     "59": "NON_PARTICIPATING",
-    "6": "SUCCESS",
-    "60": "SUCCESS",
+    "6": "NON_PARTICIPATING",
+    "60": "NON_PARTICIPATING",
     "61": "NON_PARTICIPATING",
     "62": "NON_PARTICIPATING",
-    "63": "NON_PARTICIPATING",
+    "63": "SUCCESS",
     "64": "SUCCESS",
     "65": "SUCCESS",
     "66": "SUCCESS",
     "67": "NON_PARTICIPATING",
     "68": "NON_PARTICIPATING",
-    "69": "SUCCESS",
+    "69": "NON_PARTICIPATING",
     "7": "NON_PARTICIPATING",
-    "70": "SUCCESS",
+    "70": "NON_PARTICIPATING",
     "71": "NON_PARTICIPATING",
     "72": "NON_PARTICIPATING",
     "73": "SUCCESS",
     "74": "NON_PARTICIPATING",
-    "75": "NON_PARTICIPATING",
+    "75": "SUCCESS",
     "76": "SUCCESS",
-    "77": "SUCCESS",
+    "77": "NON_PARTICIPATING",
     "78": "NON_PARTICIPATING",
-    "79": "SUCCESS",
-    "8": "SUCCESS",
-    "80": "NON_PARTICIPATING",
+    "79": "NON_PARTICIPATING",
+    "8": "NON_PARTICIPATING",
+    "80": "SUCCESS",
     "81": "NON_PARTICIPATING",
     "82": "NON_PARTICIPATING",
-    "83": "NON_PARTICIPATING",
+    "83": "SUCCESS",
     "84": "SUCCESS",
-    "85": "NON_PARTICIPATING",
+    "85": "SUCCESS",
     "86": "NON_PARTICIPATING",
     "87": "NON_PARTICIPATING",
-    "88": "SUCCESS",
-    "89": "SUCCESS",
-    "9": "NON_PARTICIPATING",
-    "90": "NON_PARTICIPATING",
-    "91": "SUCCESS",
-    "92": "NON_PARTICIPATING",
+    "88": "NON_PARTICIPATING",
+    "89": "NON_PARTICIPATING",
+    "9": "SUCCESS",
+    "90": "SUCCESS",
+    "91": "NON_PARTICIPATING",
+    "92": "SUCCESS",
     "93": "NON_PARTICIPATING",
     "94": "SUCCESS",
     "95": "NON_PARTICIPATING",
-    "96": "SUCCESS",
+    "96": "NON_PARTICIPATING",
     "97": "NON_PARTICIPATING",
     "98": "NON_PARTICIPATING",
-    "99": "NON_PARTICIPATING"
+    "99": "SUCCESS"
   },
   "architectures": [
     "GPTOptim"

model-00001-of-00002.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d8f49e3477c9013a52561eca9d73f46484083525d1538ee218ed22c4e3e6c2c1
+oid sha256:446bf25b5164f839797066ebadeaf72313e65214041231df4c598e846d38c29e
 size 4975530824

model-00002-of-00002.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f6aca6f645cc1582f97b3bfe3e7404121c4a58a7d7e441264e9b444dc2f808aa
+oid sha256:51f5bf624e9aa462beccfe387af70e7e97009a4ac6517f8033ac1b6b8070fe3b
 size 3075942944

optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:41a34596aea161c1e7a394b0cdfc5c38802f8adacc71ea57292ab82d47599b5e
+oid sha256:8ba81e34e8d229e6666e3b8a075e26f13b21cf2b849c2aa664f7d6d30b46a96d
 size 4675755870