Epoch 1155. Batch Size 44832. Peers 87.

config.json

@@ -3,65 +3,65 @@
   "activation_function": "gelu_new",
   "all_reduce_scores": {
     "0": "NON_PARTICIPATING",
-    "1": "SUCCESS",
-    "10": "SUCCESS",
+    "1": "NON_PARTICIPATING",
+    "10": "NON_PARTICIPATING",
     "100": "NON_PARTICIPATING",
     "101": "NON_PARTICIPATING",
     "102": "NON_PARTICIPATING",
-    "103": "NON_PARTICIPATING",
+    "103": "SUCCESS",
     "104": "SUCCESS",
-    "105": "NON_PARTICIPATING",
+    "105": "SUCCESS",
     "106": "SUCCESS",
-    "107": "SUCCESS",
-    "108": "SUCCESS",
-    "109": "SUCCESS",
+    "107": "NON_PARTICIPATING",
+    "108": "NON_PARTICIPATING",
+    "109": "NON_PARTICIPATING",
     "11": "NON_PARTICIPATING",
-    "110": "NON_PARTICIPATING",
+    "110": "SUCCESS",
     "111": "SUCCESS",
     "112": "NON_PARTICIPATING",
     "113": "SUCCESS",
     "114": "SUCCESS",
     "115": "SUCCESS",
-    "116": "SUCCESS",
+    "116": "NON_PARTICIPATING",
     "117": "NON_PARTICIPATING",
     "118": "NON_PARTICIPATING",
     "119": "NON_PARTICIPATING",
-    "12": "NON_PARTICIPATING",
+    "12": "SUCCESS",
     "120": "NON_PARTICIPATING",
     "121": "SUCCESS",
     "122": "NON_PARTICIPATING",
     "123": "NON_PARTICIPATING",
-    "124": "SUCCESS",
+    "124": "NON_PARTICIPATING",
     "125": "SUCCESS",
     "126": "NON_PARTICIPATING",
-    "127": "SUCCESS",
+    "127": "NON_PARTICIPATING",
     "128": "NON_PARTICIPATING",
     "129": "NON_PARTICIPATING",
-    "13": "NON_PARTICIPATING",
+    "13": "SUCCESS",
     "130": "NON_PARTICIPATING",
     "131": "NON_PARTICIPATING",
-    "132": "SUCCESS",
-    "133": "NON_PARTICIPATING",
+    "132": "NON_PARTICIPATING",
+    "133": "SUCCESS",
     "134": "NON_PARTICIPATING",
-    "135": "NON_PARTICIPATING",
+    "135": "SUCCESS",
     "136": "NON_PARTICIPATING",
     "137": "SUCCESS",
     "138": "NON_PARTICIPATING",
-    "139": "NON_PARTICIPATING",
+    "139": "SUCCESS",
     "14": "NON_PARTICIPATING",
     "140": "NON_PARTICIPATING",
     "141": "NON_PARTICIPATING",
     "142": "NON_PARTICIPATING",
-    "143": "SUCCESS",
-    "144": "NON_PARTICIPATING",
-    "145": "NON_PARTICIPATING",
+    "143": "NON_PARTICIPATING",
+    "144": "SUCCESS",
+    "145": "SUCCESS",
     "146": "NON_PARTICIPATING",
     "147": "SUCCESS",
-    "148": "NON_PARTICIPATING",
-    "149": "NON_PARTICIPATING",
-    "15": "SUCCESS",
-    "150": "NON_PARTICIPATING",
-    "151": "NON_PARTICIPATING",
+    "148": "SUCCESS",
+    "149": "SUCCESS",
+    "15": "NON_PARTICIPATING",
+    "150": "SUCCESS",
+    "151": "SUCCESS",
     "152": "NON_PARTICIPATING",
     "153": "SUCCESS",
     "154": "NON_PARTICIPATING",
@@ -69,195 +69,195 @@
     "156": "NON_PARTICIPATING",
     "157": "NON_PARTICIPATING",
     "158": "NON_PARTICIPATING",
-    "159": "NON_PARTICIPATING",
-    "16": "NON_PARTICIPATING",
+    "159": "SUCCESS",
+    "16": "SUCCESS",
     "160": "NON_PARTICIPATING",
     "161": "NON_PARTICIPATING",
-    "162": "NON_PARTICIPATING",
+    "162": "SUCCESS",
     "163": "SUCCESS",
     "164": "NON_PARTICIPATING",
     "165": "NON_PARTICIPATING",
     "166": "NON_PARTICIPATING",
-    "167": "NON_PARTICIPATING",
+    "167": "SUCCESS",
     "168": "NON_PARTICIPATING",
     "169": "NON_PARTICIPATING",
-    "17": "NON_PARTICIPATING",
+    "17": "SUCCESS",
     "170": "NON_PARTICIPATING",
-    "171": "NON_PARTICIPATING",
-    "172": "SUCCESS",
+    "171": "SUCCESS",
+    "172": "NON_PARTICIPATING",
     "173": "SUCCESS",
     "174": "SUCCESS",
-    "175": "SUCCESS",
-    "176": "NON_PARTICIPATING",
-    "177": "SUCCESS",
+    "175": "NON_PARTICIPATING",
+    "176": "SUCCESS",
+    "177": "NON_PARTICIPATING",
     "178": "NON_PARTICIPATING",
-    "179": "SUCCESS",
-    "18": "NON_PARTICIPATING",
-    "180": "SUCCESS",
+    "179": "NON_PARTICIPATING",
+    "18": "SUCCESS",
+    "180": "NON_PARTICIPATING",
     "181": "NON_PARTICIPATING",
     "182": "NON_PARTICIPATING",
     "183": "SUCCESS",
     "184": "SUCCESS",
     "185": "NON_PARTICIPATING",
     "186": "NON_PARTICIPATING",
-    "187": "SUCCESS",
+    "187": "NON_PARTICIPATING",
     "188": "NON_PARTICIPATING",
-    "189": "SUCCESS",
-    "19": "NON_PARTICIPATING",
+    "189": "NON_PARTICIPATING",
+    "19": "SUCCESS",
     "190": "NON_PARTICIPATING",
     "191": "NON_PARTICIPATING",
-    "192": "SUCCESS",
+    "192": "NON_PARTICIPATING",
     "193": "SUCCESS",
     "194": "NON_PARTICIPATING",
-    "195": "NON_PARTICIPATING",
-    "196": "SUCCESS",
+    "195": "SUCCESS",
+    "196": "NON_PARTICIPATING",
     "197": "NON_PARTICIPATING",
     "198": "NON_PARTICIPATING",
     "199": "SUCCESS",
     "2": "NON_PARTICIPATING",
-    "20": "NON_PARTICIPATING",
+    "20": "SUCCESS",
     "200": "SUCCESS",
-    "201": "NON_PARTICIPATING",
+    "201": "SUCCESS",
     "202": "NON_PARTICIPATING",
-    "203": "NON_PARTICIPATING",
-    "204": "SUCCESS",
+    "203": "SUCCESS",
+    "204": "NON_PARTICIPATING",
     "205": "NON_PARTICIPATING",
-    "206": "SUCCESS",
+    "206": "NON_PARTICIPATING",
     "207": "NON_PARTICIPATING",
-    "208": "SUCCESS",
-    "209": "NON_PARTICIPATING",
-    "21": "NON_PARTICIPATING",
-    "210": "NON_PARTICIPATING",
+    "208": "NON_PARTICIPATING",
+    "209": "SUCCESS",
+    "21": "SUCCESS",
+    "210": "SUCCESS",
     "211": "NON_PARTICIPATING",
     "212": "NON_PARTICIPATING",
-    "213": "SUCCESS",
+    "213": "NON_PARTICIPATING",
     "214": "NON_PARTICIPATING",
-    "215": "NON_PARTICIPATING",
+    "215": "SUCCESS",
     "216": "NON_PARTICIPATING",
     "217": "NON_PARTICIPATING",
     "218": "NON_PARTICIPATING",
     "219": "NON_PARTICIPATING",
     "22": "NON_PARTICIPATING",
     "220": "NON_PARTICIPATING",
-    "221": "SUCCESS",
+    "221": "NON_PARTICIPATING",
     "222": "NON_PARTICIPATING",
     "223": "NON_PARTICIPATING",
     "224": "NON_PARTICIPATING",
     "225": "NON_PARTICIPATING",
     "226": "NON_PARTICIPATING",
     "227": "SUCCESS",
-    "228": "SUCCESS",
+    "228": "NON_PARTICIPATING",
     "229": "NON_PARTICIPATING",
-    "23": "NON_PARTICIPATING",
+    "23": "SUCCESS",
     "230": "NON_PARTICIPATING",
     "231": "SUCCESS",
-    "232": "SUCCESS",
+    "232": "NON_PARTICIPATING",
     "233": "NON_PARTICIPATING",
     "234": "NON_PARTICIPATING",
     "235": "SUCCESS",
-    "236": "SUCCESS",
+    "236": "NON_PARTICIPATING",
     "237": "NON_PARTICIPATING",
-    "238": "SUCCESS",
-    "239": "NON_PARTICIPATING",
-    "24": "SUCCESS",
-    "240": "NON_PARTICIPATING",
+    "238": "NON_PARTICIPATING",
+    "239": "SUCCESS",
+    "24": "NON_PARTICIPATING",
+    "240": "SUCCESS",
     "241": "NON_PARTICIPATING",
     "242": "SUCCESS",
-    "243": "SUCCESS",
-    "244": "SUCCESS",
-    "245": "SUCCESS",
+    "243": "NON_PARTICIPATING",
+    "244": "NON_PARTICIPATING",
+    "245": "NON_PARTICIPATING",
     "246": "NON_PARTICIPATING",
     "247": "NON_PARTICIPATING",
-    "248": "SUCCESS",
+    "248": "NON_PARTICIPATING",
     "249": "NON_PARTICIPATING",
     "25": "NON_PARTICIPATING",
     "250": "NON_PARTICIPATING",
-    "251": "SUCCESS",
+    "251": "NON_PARTICIPATING",
     "252": "NON_PARTICIPATING",
     "253": "NON_PARTICIPATING",
     "254": "NON_PARTICIPATING",
-    "255": "SUCCESS",
-    "26": "SUCCESS",
-    "27": "SUCCESS",
-    "28": "SUCCESS",
+    "255": "NON_PARTICIPATING",
+    "26": "NON_PARTICIPATING",
+    "27": "NON_PARTICIPATING",
+    "28": "NON_PARTICIPATING",
     "29": "NON_PARTICIPATING",
-    "3": "NON_PARTICIPATING",
-    "30": "SUCCESS",
+    "3": "SUCCESS",
+    "30": "NON_PARTICIPATING",
     "31": "SUCCESS",
     "32": "NON_PARTICIPATING",
     "33": "NON_PARTICIPATING",
-    "34": "NON_PARTICIPATING",
+    "34": "SUCCESS",
     "35": "NON_PARTICIPATING",
     "36": "NON_PARTICIPATING",
     "37": "NON_PARTICIPATING",
     "38": "NON_PARTICIPATING",
-    "39": "NON_PARTICIPATING",
+    "39": "SUCCESS",
     "4": "NON_PARTICIPATING",
     "40": "NON_PARTICIPATING",
-    "41": "SUCCESS",
-    "42": "SUCCESS",
+    "41": "NON_PARTICIPATING",
+    "42": "NON_PARTICIPATING",
     "43": "SUCCESS",
-    "44": "NON_PARTICIPATING",
+    "44": "SUCCESS",
     "45": "NON_PARTICIPATING",
     "46": "NON_PARTICIPATING",
     "47": "SUCCESS",
-    "48": "SUCCESS",
+    "48": "NON_PARTICIPATING",
     "49": "NON_PARTICIPATING",
     "5": "NON_PARTICIPATING",
     "50": "SUCCESS",
     "51": "NON_PARTICIPATING",
-    "52": "SUCCESS",
+    "52": "NON_PARTICIPATING",
     "53": "NON_PARTICIPATING",
-    "54": "NON_PARTICIPATING",
+    "54": "SUCCESS",
     "55": "SUCCESS",
     "56": "NON_PARTICIPATING",
     "57": "NON_PARTICIPATING",
     "58": "NON_PARTICIPATING",
     "59": "NON_PARTICIPATING",
-    "6": "NON_PARTICIPATING",
+    "6": "SUCCESS",
     "60": "SUCCESS",
-    "61": "SUCCESS",
+    "61": "NON_PARTICIPATING",
     "62": "NON_PARTICIPATING",
     "63": "SUCCESS",
     "64": "NON_PARTICIPATING",
     "65": "SUCCESS",
-    "66": "NON_PARTICIPATING",
+    "66": "SUCCESS",
     "67": "NON_PARTICIPATING",
-    "68": "NON_PARTICIPATING",
+    "68": "SUCCESS",
     "69": "NON_PARTICIPATING",
-    "7": "NON_PARTICIPATING",
-    "70": "NON_PARTICIPATING",
+    "7": "SUCCESS",
+    "70": "SUCCESS",
     "71": "SUCCESS",
     "72": "NON_PARTICIPATING",
     "73": "NON_PARTICIPATING",
-    "74": "SUCCESS",
+    "74": "NON_PARTICIPATING",
     "75": "NON_PARTICIPATING",
-    "76": "NON_PARTICIPATING",
+    "76": "SUCCESS",
     "77": "NON_PARTICIPATING",
-    "78": "NON_PARTICIPATING",
-    "79": "NON_PARTICIPATING",
+    "78": "SUCCESS",
+    "79": "SUCCESS",
     "8": "NON_PARTICIPATING",
     "80": "NON_PARTICIPATING",
-    "81": "SUCCESS",
+    "81": "NON_PARTICIPATING",
     "82": "NON_PARTICIPATING",
-    "83": "NON_PARTICIPATING",
-    "84": "SUCCESS",
-    "85": "SUCCESS",
-    "86": "SUCCESS",
+    "83": "SUCCESS",
+    "84": "NON_PARTICIPATING",
+    "85": "NON_PARTICIPATING",
+    "86": "NON_PARTICIPATING",
     "87": "NON_PARTICIPATING",
     "88": "SUCCESS",
-    "89": "NON_PARTICIPATING",
-    "9": "SUCCESS",
-    "90": "NON_PARTICIPATING",
+    "89": "SUCCESS",
+    "9": "NON_PARTICIPATING",
+    "90": "SUCCESS",
     "91": "SUCCESS",
     "92": "NON_PARTICIPATING",
-    "93": "SUCCESS",
+    "93": "NON_PARTICIPATING",
     "94": "NON_PARTICIPATING",
     "95": "NON_PARTICIPATING",
     "96": "NON_PARTICIPATING",
     "97": "NON_PARTICIPATING",
-    "98": "NON_PARTICIPATING",
-    "99": "NON_PARTICIPATING"
+    "98": "SUCCESS",
+    "99": "SUCCESS"
   },
   "architectures": [
     "GPTOptim"

model-00001-of-00002.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:35335ee1046cfb8208549cd952140c9b2075dcc2228bffae82b56ae6b1996108
+oid sha256:e7f4ba5dc64c8c1eab3ffc02697d8ec74481a9f996a3821bb30ccfbb5db9ca06
 size 4975530824

model-00002-of-00002.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f02c6aa2fbeae71907b16b56c1ae8f0a20677c90969189df4ac64a612781c9f3
+oid sha256:db0927b60f3de17d4c08f0ec42d610f69f2b61151c3502f6a68adc8f0b204adc
 size 3075942944

optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a5bc709cf2f0a1f062b6a807ac159246728a7b043e0b335ba2f1953697c4d9fb
+oid sha256:a3573a1ef9ce679656ece0c098e48775de6e59febe6796bada867c66e75f7c39
 size 4675755870