Epoch 850. Batch Size 40272. Peers 85.

config.json

@@ -3,259 +3,259 @@
   "activation_function": "gelu_new",
   "all_reduce_scores": {
     "0": "NON_PARTICIPATING",
-    "1": "NON_PARTICIPATING",
-    "10": "NON_PARTICIPATING",
+    "1": "SUCCESS",
+    "10": "SUCCESS",
     "100": "NON_PARTICIPATING",
-    "101": "SUCCESS",
+    "101": "NON_PARTICIPATING",
     "102": "NON_PARTICIPATING",
     "103": "NON_PARTICIPATING",
-    "104": "NON_PARTICIPATING",
+    "104": "SUCCESS",
     "105": "SUCCESS",
     "106": "NON_PARTICIPATING",
-    "107": "SUCCESS",
+    "107": "NON_PARTICIPATING",
     "108": "NON_PARTICIPATING",
-    "109": "SUCCESS",
+    "109": "NON_PARTICIPATING",
     "11": "NON_PARTICIPATING",
     "110": "SUCCESS",
     "111": "NON_PARTICIPATING",
-    "112": "SUCCESS",
-    "113": "SUCCESS",
+    "112": "NON_PARTICIPATING",
+    "113": "NON_PARTICIPATING",
     "114": "SUCCESS",
     "115": "NON_PARTICIPATING",
-    "116": "SUCCESS",
+    "116": "NON_PARTICIPATING",
     "117": "NON_PARTICIPATING",
-    "118": "NON_PARTICIPATING",
-    "119": "NON_PARTICIPATING",
+    "118": "SUCCESS",
+    "119": "SUCCESS",
     "12": "NON_PARTICIPATING",
     "120": "NON_PARTICIPATING",
     "121": "NON_PARTICIPATING",
-    "122": "NON_PARTICIPATING",
+    "122": "SUCCESS",
     "123": "NON_PARTICIPATING",
-    "124": "SUCCESS",
+    "124": "NON_PARTICIPATING",
     "125": "NON_PARTICIPATING",
     "126": "NON_PARTICIPATING",
     "127": "SUCCESS",
     "128": "SUCCESS",
     "129": "SUCCESS",
     "13": "NON_PARTICIPATING",
-    "130": "SUCCESS",
+    "130": "NON_PARTICIPATING",
     "131": "NON_PARTICIPATING",
-    "132": "SUCCESS",
-    "133": "NON_PARTICIPATING",
-    "134": "SUCCESS",
+    "132": "NON_PARTICIPATING",
+    "133": "SUCCESS",
+    "134": "NON_PARTICIPATING",
     "135": "NON_PARTICIPATING",
     "136": "NON_PARTICIPATING",
-    "137": "NON_PARTICIPATING",
+    "137": "SUCCESS",
     "138": "NON_PARTICIPATING",
     "139": "NON_PARTICIPATING",
-    "14": "SUCCESS",
+    "14": "NON_PARTICIPATING",
     "140": "NON_PARTICIPATING",
-    "141": "NON_PARTICIPATING",
-    "142": "NON_PARTICIPATING",
+    "141": "SUCCESS",
+    "142": "SUCCESS",
     "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",
-    "150": "NON_PARTICIPATING",
-    "151": "SUCCESS",
-    "152": "NON_PARTICIPATING",
+    "150": "SUCCESS",
+    "151": "NON_PARTICIPATING",
+    "152": "SUCCESS",
     "153": "NON_PARTICIPATING",
-    "154": "SUCCESS",
-    "155": "SUCCESS",
+    "154": "NON_PARTICIPATING",
+    "155": "NON_PARTICIPATING",
     "156": "NON_PARTICIPATING",
     "157": "SUCCESS",
     "158": "NON_PARTICIPATING",
-    "159": "SUCCESS",
-    "16": "SUCCESS",
+    "159": "NON_PARTICIPATING",
+    "16": "NON_PARTICIPATING",
     "160": "NON_PARTICIPATING",
     "161": "NON_PARTICIPATING",
-    "162": "SUCCESS",
+    "162": "NON_PARTICIPATING",
     "163": "NON_PARTICIPATING",
-    "164": "SUCCESS",
-    "165": "SUCCESS",
-    "166": "NON_PARTICIPATING",
+    "164": "NON_PARTICIPATING",
+    "165": "NON_PARTICIPATING",
+    "166": "SUCCESS",
     "167": "NON_PARTICIPATING",
     "168": "NON_PARTICIPATING",
     "169": "NON_PARTICIPATING",
     "17": "NON_PARTICIPATING",
     "170": "NON_PARTICIPATING",
-    "171": "NON_PARTICIPATING",
+    "171": "SUCCESS",
     "172": "NON_PARTICIPATING",
-    "173": "NON_PARTICIPATING",
-    "174": "NON_PARTICIPATING",
+    "173": "SUCCESS",
+    "174": "SUCCESS",
     "175": "SUCCESS",
     "176": "SUCCESS",
     "177": "NON_PARTICIPATING",
     "178": "SUCCESS",
     "179": "NON_PARTICIPATING",
     "18": "NON_PARTICIPATING",
-    "180": "SUCCESS",
-    "181": "SUCCESS",
+    "180": "NON_PARTICIPATING",
+    "181": "NON_PARTICIPATING",
     "182": "NON_PARTICIPATING",
     "183": "NON_PARTICIPATING",
-    "184": "NON_PARTICIPATING",
+    "184": "SUCCESS",
     "185": "NON_PARTICIPATING",
-    "186": "SUCCESS",
+    "186": "NON_PARTICIPATING",
     "187": "SUCCESS",
-    "188": "NON_PARTICIPATING",
-    "189": "NON_PARTICIPATING",
-    "19": "SUCCESS",
+    "188": "SUCCESS",
+    "189": "SUCCESS",
+    "19": "NON_PARTICIPATING",
     "190": "NON_PARTICIPATING",
     "191": "NON_PARTICIPATING",
     "192": "NON_PARTICIPATING",
     "193": "SUCCESS",
     "194": "NON_PARTICIPATING",
-    "195": "SUCCESS",
-    "196": "SUCCESS",
-    "197": "SUCCESS",
+    "195": "NON_PARTICIPATING",
+    "196": "NON_PARTICIPATING",
+    "197": "NON_PARTICIPATING",
     "198": "SUCCESS",
     "199": "NON_PARTICIPATING",
     "2": "NON_PARTICIPATING",
     "20": "NON_PARTICIPATING",
-    "200": "NON_PARTICIPATING",
-    "201": "SUCCESS",
+    "200": "SUCCESS",
+    "201": "NON_PARTICIPATING",
     "202": "NON_PARTICIPATING",
-    "203": "SUCCESS",
+    "203": "NON_PARTICIPATING",
     "204": "NON_PARTICIPATING",
     "205": "NON_PARTICIPATING",
     "206": "NON_PARTICIPATING",
     "207": "NON_PARTICIPATING",
     "208": "SUCCESS",
     "209": "SUCCESS",
-    "21": "SUCCESS",
-    "210": "NON_PARTICIPATING",
-    "211": "NON_PARTICIPATING",
-    "212": "SUCCESS",
+    "21": "NON_PARTICIPATING",
+    "210": "SUCCESS",
+    "211": "SUCCESS",
+    "212": "NON_PARTICIPATING",
     "213": "NON_PARTICIPATING",
-    "214": "NON_PARTICIPATING",
+    "214": "SUCCESS",
     "215": "SUCCESS",
-    "216": "SUCCESS",
+    "216": "NON_PARTICIPATING",
     "217": "SUCCESS",
     "218": "SUCCESS",
     "219": "NON_PARTICIPATING",
-    "22": "NON_PARTICIPATING",
+    "22": "SUCCESS",
     "220": "NON_PARTICIPATING",
-    "221": "NON_PARTICIPATING",
+    "221": "SUCCESS",
     "222": "NON_PARTICIPATING",
     "223": "NON_PARTICIPATING",
     "224": "NON_PARTICIPATING",
-    "225": "NON_PARTICIPATING",
+    "225": "SUCCESS",
     "226": "NON_PARTICIPATING",
-    "227": "SUCCESS",
+    "227": "NON_PARTICIPATING",
     "228": "NON_PARTICIPATING",
     "229": "SUCCESS",
-    "23": "SUCCESS",
+    "23": "NON_PARTICIPATING",
     "230": "NON_PARTICIPATING",
     "231": "NON_PARTICIPATING",
     "232": "NON_PARTICIPATING",
-    "233": "SUCCESS",
+    "233": "NON_PARTICIPATING",
     "234": "NON_PARTICIPATING",
     "235": "NON_PARTICIPATING",
     "236": "NON_PARTICIPATING",
-    "237": "NON_PARTICIPATING",
-    "238": "NON_PARTICIPATING",
+    "237": "SUCCESS",
+    "238": "SUCCESS",
     "239": "NON_PARTICIPATING",
     "24": "NON_PARTICIPATING",
     "240": "NON_PARTICIPATING",
     "241": "SUCCESS",
     "242": "SUCCESS",
     "243": "SUCCESS",
-    "244": "NON_PARTICIPATING",
-    "245": "SUCCESS",
-    "246": "SUCCESS",
-    "247": "NON_PARTICIPATING",
-    "248": "SUCCESS",
+    "244": "SUCCESS",
+    "245": "NON_PARTICIPATING",
+    "246": "NON_PARTICIPATING",
+    "247": "SUCCESS",
+    "248": "NON_PARTICIPATING",
     "249": "NON_PARTICIPATING",
     "25": "SUCCESS",
     "250": "SUCCESS",
     "251": "NON_PARTICIPATING",
-    "252": "SUCCESS",
-    "253": "SUCCESS",
+    "252": "NON_PARTICIPATING",
+    "253": "NON_PARTICIPATING",
     "254": "NON_PARTICIPATING",
     "255": "NON_PARTICIPATING",
     "26": "NON_PARTICIPATING",
     "27": "NON_PARTICIPATING",
-    "28": "NON_PARTICIPATING",
-    "29": "SUCCESS",
-    "3": "SUCCESS",
-    "30": "NON_PARTICIPATING",
-    "31": "SUCCESS",
+    "28": "SUCCESS",
+    "29": "NON_PARTICIPATING",
+    "3": "NON_PARTICIPATING",
+    "30": "SUCCESS",
+    "31": "NON_PARTICIPATING",
     "32": "NON_PARTICIPATING",
     "33": "NON_PARTICIPATING",
     "34": "NON_PARTICIPATING",
-    "35": "NON_PARTICIPATING",
-    "36": "SUCCESS",
+    "35": "SUCCESS",
+    "36": "NON_PARTICIPATING",
     "37": "NON_PARTICIPATING",
     "38": "NON_PARTICIPATING",
-    "39": "SUCCESS",
-    "4": "SUCCESS",
+    "39": "NON_PARTICIPATING",
+    "4": "NON_PARTICIPATING",
     "40": "NON_PARTICIPATING",
-    "41": "NON_PARTICIPATING",
-    "42": "NON_PARTICIPATING",
+    "41": "SUCCESS",
+    "42": "SUCCESS",
     "43": "NON_PARTICIPATING",
-    "44": "NON_PARTICIPATING",
-    "45": "SUCCESS",
+    "44": "SUCCESS",
+    "45": "NON_PARTICIPATING",
     "46": "NON_PARTICIPATING",
-    "47": "NON_PARTICIPATING",
+    "47": "SUCCESS",
     "48": "NON_PARTICIPATING",
-    "49": "NON_PARTICIPATING",
+    "49": "SUCCESS",
     "5": "NON_PARTICIPATING",
     "50": "NON_PARTICIPATING",
     "51": "NON_PARTICIPATING",
-    "52": "NON_PARTICIPATING",
-    "53": "NON_PARTICIPATING",
+    "52": "SUCCESS",
+    "53": "SUCCESS",
     "54": "SUCCESS",
     "55": "NON_PARTICIPATING",
-    "56": "SUCCESS",
+    "56": "NON_PARTICIPATING",
     "57": "NON_PARTICIPATING",
     "58": "NON_PARTICIPATING",
-    "59": "SUCCESS",
-    "6": "NON_PARTICIPATING",
-    "60": "SUCCESS",
+    "59": "NON_PARTICIPATING",
+    "6": "SUCCESS",
+    "60": "NON_PARTICIPATING",
     "61": "NON_PARTICIPATING",
     "62": "SUCCESS",
-    "63": "NON_PARTICIPATING",
-    "64": "SUCCESS",
-    "65": "SUCCESS",
+    "63": "SUCCESS",
+    "64": "NON_PARTICIPATING",
+    "65": "NON_PARTICIPATING",
     "66": "NON_PARTICIPATING",
-    "67": "SUCCESS",
+    "67": "NON_PARTICIPATING",
     "68": "NON_PARTICIPATING",
-    "69": "NON_PARTICIPATING",
+    "69": "SUCCESS",
     "7": "SUCCESS",
-    "70": "NON_PARTICIPATING",
+    "70": "SUCCESS",
     "71": "NON_PARTICIPATING",
     "72": "NON_PARTICIPATING",
     "73": "NON_PARTICIPATING",
-    "74": "SUCCESS",
+    "74": "NON_PARTICIPATING",
     "75": "SUCCESS",
     "76": "SUCCESS",
-    "77": "NON_PARTICIPATING",
-    "78": "NON_PARTICIPATING",
+    "77": "SUCCESS",
+    "78": "SUCCESS",
     "79": "NON_PARTICIPATING",
     "8": "SUCCESS",
     "80": "NON_PARTICIPATING",
-    "81": "SUCCESS",
-    "82": "SUCCESS",
+    "81": "NON_PARTICIPATING",
+    "82": "NON_PARTICIPATING",
     "83": "SUCCESS",
     "84": "NON_PARTICIPATING",
     "85": "NON_PARTICIPATING",
-    "86": "NON_PARTICIPATING",
-    "87": "SUCCESS",
+    "86": "SUCCESS",
+    "87": "NON_PARTICIPATING",
     "88": "NON_PARTICIPATING",
     "89": "NON_PARTICIPATING",
     "9": "NON_PARTICIPATING",
     "90": "NON_PARTICIPATING",
-    "91": "SUCCESS",
-    "92": "NON_PARTICIPATING",
-    "93": "NON_PARTICIPATING",
+    "91": "NON_PARTICIPATING",
+    "92": "SUCCESS",
+    "93": "SUCCESS",
     "94": "SUCCESS",
     "95": "NON_PARTICIPATING",
     "96": "SUCCESS",
-    "97": "NON_PARTICIPATING",
+    "97": "SUCCESS",
     "98": "NON_PARTICIPATING",
     "99": "NON_PARTICIPATING"
   },

model-00001-of-00002.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:14a87314ac793f1fbc6441cf16cc77d571f8ed87fe97887c682916652e46226f
+oid sha256:324f722c2501277887e60e392070d63ce88d393aa6f436ecd134ad45a256ab75
 size 4975530824

model-00002-of-00002.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:1f402d2e376f7ffc9a734dcb12b8cfb1bb00d2c9dac46402dbaed20a0ff396e6
+oid sha256:ed91212c83a6a01a17f29654ae31b03fefbdf5aaa3b962db0ae6e35fe2be50bb
 size 3075942944

optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ba655908c1402a23c6857061f4212f7a42d03980ae538dd04a62240f973b91b2
+oid sha256:ee589c81463d3695e3c8ecaf3c73da73e96ac9f2b7bd9413a61927347d77f417
 size 4675755870