Run 4. Outer Step 51. Inner Step 2.

config.json

@@ -1,9 +1,9 @@
 {
-  "_name_or_path": "philip324mark/mark16",
+  "_name_or_path": "distributed/optimized-gpt2-1b",
   "activation_function": "gelu_new",
   "all_reduce_scores": {
     "0": "NON_PARTICIPATING",
-    "1": "SUCCESS",
+    "1": "NON_PARTICIPATING",
     "10": "NON_PARTICIPATING",
     "100": "NON_PARTICIPATING",
     "101": "NON_PARTICIPATING",
@@ -12,27 +12,27 @@
     "104": "NON_PARTICIPATING",
     "105": "NON_PARTICIPATING",
     "106": "NON_PARTICIPATING",
-    "107": "NON_PARTICIPATING",
-    "108": "NON_PARTICIPATING",
+    "107": "SUCCESS",
+    "108": "SUCCESS",
     "109": "NON_PARTICIPATING",
     "11": "SUCCESS",
-    "110": "NON_PARTICIPATING",
+    "110": "SUCCESS",
     "111": "NON_PARTICIPATING",
     "112": "NON_PARTICIPATING",
     "113": "NON_PARTICIPATING",
     "114": "NON_PARTICIPATING",
     "115": "NON_PARTICIPATING",
     "116": "NON_PARTICIPATING",
-    "117": "NON_PARTICIPATING",
-    "118": "SUCCESS",
+    "117": "SUCCESS",
+    "118": "NON_PARTICIPATING",
     "119": "NON_PARTICIPATING",
     "12": "SUCCESS",
     "120": "NON_PARTICIPATING",
     "121": "NON_PARTICIPATING",
-    "122": "NON_PARTICIPATING",
+    "122": "SUCCESS",
     "123": "NON_PARTICIPATING",
     "124": "NON_PARTICIPATING",
-    "125": "NON_PARTICIPATING",
+    "125": "SUCCESS",
     "126": "NON_PARTICIPATING",
     "127": "NON_PARTICIPATING",
     "128": "NON_PARTICIPATING",
@@ -42,16 +42,16 @@
     "131": "NON_PARTICIPATING",
     "132": "NON_PARTICIPATING",
     "133": "NON_PARTICIPATING",
-    "134": "NON_PARTICIPATING",
+    "134": "SUCCESS",
     "135": "NON_PARTICIPATING",
-    "136": "SUCCESS",
+    "136": "NON_PARTICIPATING",
     "137": "NON_PARTICIPATING",
-    "138": "SUCCESS",
+    "138": "NON_PARTICIPATING",
     "139": "NON_PARTICIPATING",
-    "14": "SUCCESS",
+    "14": "NON_PARTICIPATING",
     "140": "NON_PARTICIPATING",
-    "141": "NON_PARTICIPATING",
-    "142": "SUCCESS",
+    "141": "SUCCESS",
+    "142": "NON_PARTICIPATING",
     "143": "NON_PARTICIPATING",
     "144": "NON_PARTICIPATING",
     "145": "NON_PARTICIPATING",
@@ -60,18 +60,18 @@
     "148": "NON_PARTICIPATING",
     "149": "NON_PARTICIPATING",
     "15": "NON_PARTICIPATING",
-    "150": "NON_PARTICIPATING",
+    "150": "SUCCESS",
     "151": "NON_PARTICIPATING",
     "152": "NON_PARTICIPATING",
     "153": "NON_PARTICIPATING",
     "154": "NON_PARTICIPATING",
     "155": "NON_PARTICIPATING",
-    "156": "NON_PARTICIPATING",
-    "157": "SUCCESS",
+    "156": "SUCCESS",
+    "157": "NON_PARTICIPATING",
     "158": "NON_PARTICIPATING",
     "159": "NON_PARTICIPATING",
     "16": "SUCCESS",
-    "160": "SUCCESS",
+    "160": "NON_PARTICIPATING",
     "161": "NON_PARTICIPATING",
     "162": "NON_PARTICIPATING",
     "163": "NON_PARTICIPATING",
@@ -87,10 +87,10 @@
     "172": "NON_PARTICIPATING",
     "173": "NON_PARTICIPATING",
     "174": "NON_PARTICIPATING",
-    "175": "SUCCESS",
+    "175": "NON_PARTICIPATING",
     "176": "NON_PARTICIPATING",
-    "177": "SUCCESS",
-    "178": "NON_PARTICIPATING",
+    "177": "NON_PARTICIPATING",
+    "178": "SUCCESS",
     "179": "NON_PARTICIPATING",
     "18": "SUCCESS",
     "180": "NON_PARTICIPATING",
@@ -100,10 +100,10 @@
     "184": "NON_PARTICIPATING",
     "185": "NON_PARTICIPATING",
     "186": "NON_PARTICIPATING",
-    "187": "NON_PARTICIPATING",
+    "187": "SUCCESS",
     "188": "NON_PARTICIPATING",
     "189": "NON_PARTICIPATING",
-    "19": "SUCCESS",
+    "19": "NON_PARTICIPATING",
     "190": "NON_PARTICIPATING",
     "191": "NON_PARTICIPATING",
     "192": "NON_PARTICIPATING",
@@ -117,9 +117,9 @@
     "2": "SUCCESS",
     "20": "SUCCESS",
     "200": "NON_PARTICIPATING",
-    "201": "SUCCESS",
+    "201": "NON_PARTICIPATING",
     "202": "NON_PARTICIPATING",
-    "203": "SUCCESS",
+    "203": "NON_PARTICIPATING",
     "204": "NON_PARTICIPATING",
     "205": "NON_PARTICIPATING",
     "206": "NON_PARTICIPATING",
@@ -132,42 +132,42 @@
     "212": "NON_PARTICIPATING",
     "213": "NON_PARTICIPATING",
     "214": "NON_PARTICIPATING",
-    "215": "SUCCESS",
-    "216": "NON_PARTICIPATING",
+    "215": "NON_PARTICIPATING",
+    "216": "SUCCESS",
     "217": "NON_PARTICIPATING",
-    "218": "NON_PARTICIPATING",
-    "219": "NON_PARTICIPATING",
+    "218": "SUCCESS",
+    "219": "SUCCESS",
     "22": "SUCCESS",
-    "220": "SUCCESS",
+    "220": "NON_PARTICIPATING",
     "221": "NON_PARTICIPATING",
-    "222": "NON_PARTICIPATING",
+    "222": "SUCCESS",
     "223": "NON_PARTICIPATING",
     "224": "NON_PARTICIPATING",
     "225": "NON_PARTICIPATING",
     "226": "NON_PARTICIPATING",
     "227": "NON_PARTICIPATING",
     "228": "NON_PARTICIPATING",
-    "229": "SUCCESS",
+    "229": "NON_PARTICIPATING",
     "23": "SUCCESS",
     "230": "NON_PARTICIPATING",
     "231": "NON_PARTICIPATING",
     "232": "NON_PARTICIPATING",
-    "233": "SUCCESS",
+    "233": "NON_PARTICIPATING",
     "234": "NON_PARTICIPATING",
     "235": "NON_PARTICIPATING",
-    "236": "NON_PARTICIPATING",
+    "236": "SUCCESS",
     "237": "NON_PARTICIPATING",
     "238": "NON_PARTICIPATING",
     "239": "NON_PARTICIPATING",
     "24": "SUCCESS",
     "240": "NON_PARTICIPATING",
     "241": "NON_PARTICIPATING",
-    "242": "SUCCESS",
+    "242": "NON_PARTICIPATING",
     "243": "NON_PARTICIPATING",
     "244": "NON_PARTICIPATING",
     "245": "NON_PARTICIPATING",
     "246": "NON_PARTICIPATING",
-    "247": "SUCCESS",
+    "247": "NON_PARTICIPATING",
     "248": "NON_PARTICIPATING",
     "249": "NON_PARTICIPATING",
     "25": "SUCCESS",
@@ -182,20 +182,20 @@
     "28": "SUCCESS",
     "29": "SUCCESS",
     "3": "SUCCESS",
-    "30": "SUCCESS",
+    "30": "NON_PARTICIPATING",
     "31": "NON_PARTICIPATING",
     "32": "SUCCESS",
     "33": "NON_PARTICIPATING",
-    "34": "SUCCESS",
+    "34": "NON_PARTICIPATING",
     "35": "SUCCESS",
     "36": "SUCCESS",
     "37": "SUCCESS",
-    "38": "NON_PARTICIPATING",
+    "38": "SUCCESS",
     "39": "SUCCESS",
     "4": "SUCCESS",
-    "40": "SUCCESS",
+    "40": "NON_PARTICIPATING",
     "41": "SUCCESS",
-    "42": "NON_PARTICIPATING",
+    "42": "SUCCESS",
     "43": "SUCCESS",
     "44": "SUCCESS",
     "45": "SUCCESS",
@@ -206,50 +206,50 @@
     "5": "SUCCESS",
     "50": "SUCCESS",
     "51": "SUCCESS",
-    "52": "SUCCESS",
+    "52": "NON_PARTICIPATING",
     "53": "SUCCESS",
     "54": "SUCCESS",
     "55": "SUCCESS",
     "56": "SUCCESS",
-    "57": "SUCCESS",
+    "57": "NON_PARTICIPATING",
     "58": "NON_PARTICIPATING",
-    "59": "SUCCESS",
+    "59": "NON_PARTICIPATING",
     "6": "NON_PARTICIPATING",
-    "60": "SUCCESS",
+    "60": "NON_PARTICIPATING",
     "61": "NON_PARTICIPATING",
     "62": "SUCCESS",
     "63": "SUCCESS",
-    "64": "NON_PARTICIPATING",
-    "65": "SUCCESS",
+    "64": "SUCCESS",
+    "65": "NON_PARTICIPATING",
     "66": "SUCCESS",
     "67": "SUCCESS",
     "68": "SUCCESS",
     "69": "SUCCESS",
     "7": "SUCCESS",
-    "70": "NON_PARTICIPATING",
+    "70": "SUCCESS",
     "71": "NON_PARTICIPATING",
     "72": "SUCCESS",
     "73": "NON_PARTICIPATING",
     "74": "SUCCESS",
     "75": "SUCCESS",
-    "76": "SUCCESS",
-    "77": "SUCCESS",
+    "76": "NON_PARTICIPATING",
+    "77": "NON_PARTICIPATING",
     "78": "SUCCESS",
     "79": "NON_PARTICIPATING",
-    "8": "SUCCESS",
+    "8": "NON_PARTICIPATING",
     "80": "SUCCESS",
     "81": "SUCCESS",
     "82": "SUCCESS",
     "83": "NON_PARTICIPATING",
     "84": "SUCCESS",
-    "85": "NON_PARTICIPATING",
+    "85": "SUCCESS",
     "86": "NON_PARTICIPATING",
     "87": "SUCCESS",
     "88": "SUCCESS",
     "89": "SUCCESS",
     "9": "SUCCESS",
     "90": "SUCCESS",
-    "91": "SUCCESS",
+    "91": "NON_PARTICIPATING",
     "92": "SUCCESS",
     "93": "NON_PARTICIPATING",
     "94": "SUCCESS",
@@ -268,17 +268,17 @@
     "AutoModelForCausalLM": "distributed/optimized-gpt2-500m--modeling_gpt_optimized.GPTOptim"
   },
   "block_list": [
-    5626926,
-    5626954
+    5627136,
+    5627164
   ],
   "block_size": 1024,
   "bos_token_id": 50256,
   "embd_pdrop": 0.1,
   "eos_token_id": 50256,
   "initializer_range": 0.02,
-  "inner_step": 172,
+  "inner_step": 2,
   "inner_steps": 0,
-  "last_allreduce_block": 5571611,
+  "last_allreduce_block": 5627065,
   "layer_norm_epsilon": 1e-05,
   "model_type": "gpt_optimized",
   "n_embd": 1280,

inner_optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:8f3f9f536212cbf7f884c7334d772656f25a6250c92377f1fa0cfb4a4a272168
+oid sha256:539659be26453681362aaf491ddec303313beb66e8084ab0c11a848db907b069
 size 8081782503

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:716d8d4c26b0dcf39b80e1ca0f2a37b68ca6474d59d694822df7e732660ab09d
+oid sha256:16ab0d1848f982aa2b4dc4f53e679f0c12ba4a06daa23c082483dca78505f980
 size 4040701744