Run 4. Outer Step 59. Inner Step 0.

config.json

@@ -3,35 +3,35 @@
   "activation_function": "gelu_new",
   "all_reduce_scores": {
     "0": "NON_PARTICIPATING",
-    "1": "SUCCESS",
+    "1": "NON_PARTICIPATING",
     "10": "NON_PARTICIPATING",
     "100": "NON_PARTICIPATING",
-    "101": "SUCCESS",
-    "102": "NON_PARTICIPATING",
+    "101": "NON_PARTICIPATING",
+    "102": "SUCCESS",
     "103": "NON_PARTICIPATING",
-    "104": "NON_PARTICIPATING",
+    "104": "SUCCESS",
     "105": "NON_PARTICIPATING",
-    "106": "NON_PARTICIPATING",
-    "107": "SUCCESS",
-    "108": "SUCCESS",
-    "109": "NON_PARTICIPATING",
+    "106": "SUCCESS",
+    "107": "NON_PARTICIPATING",
+    "108": "NON_PARTICIPATING",
+    "109": "SUCCESS",
     "11": "SUCCESS",
-    "110": "SUCCESS",
-    "111": "NON_PARTICIPATING",
+    "110": "NON_PARTICIPATING",
+    "111": "SUCCESS",
     "112": "NON_PARTICIPATING",
     "113": "SUCCESS",
     "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": "SUCCESS",
+    "122": "NON_PARTICIPATING",
     "123": "NON_PARTICIPATING",
-    "124": "SUCCESS",
+    "124": "NON_PARTICIPATING",
     "125": "NON_PARTICIPATING",
     "126": "NON_PARTICIPATING",
     "127": "NON_PARTICIPATING",
@@ -41,7 +41,7 @@
     "130": "SUCCESS",
     "131": "NON_PARTICIPATING",
     "132": "NON_PARTICIPATING",
-    "133": "SUCCESS",
+    "133": "NON_PARTICIPATING",
     "134": "NON_PARTICIPATING",
     "135": "NON_PARTICIPATING",
     "136": "NON_PARTICIPATING",
@@ -56,17 +56,17 @@
     "144": "NON_PARTICIPATING",
     "145": "NON_PARTICIPATING",
     "146": "NON_PARTICIPATING",
-    "147": "SUCCESS",
-    "148": "NON_PARTICIPATING",
+    "147": "NON_PARTICIPATING",
+    "148": "SUCCESS",
     "149": "NON_PARTICIPATING",
     "15": "SUCCESS",
     "150": "NON_PARTICIPATING",
     "151": "NON_PARTICIPATING",
     "152": "NON_PARTICIPATING",
     "153": "NON_PARTICIPATING",
-    "154": "NON_PARTICIPATING",
+    "154": "SUCCESS",
     "155": "NON_PARTICIPATING",
-    "156": "NON_PARTICIPATING",
+    "156": "SUCCESS",
     "157": "NON_PARTICIPATING",
     "158": "NON_PARTICIPATING",
     "159": "NON_PARTICIPATING",
@@ -81,7 +81,7 @@
     "167": "NON_PARTICIPATING",
     "168": "NON_PARTICIPATING",
     "169": "NON_PARTICIPATING",
-    "17": "SUCCESS",
+    "17": "NON_PARTICIPATING",
     "170": "NON_PARTICIPATING",
     "171": "SUCCESS",
     "172": "NON_PARTICIPATING",
@@ -94,17 +94,17 @@
     "179": "NON_PARTICIPATING",
     "18": "SUCCESS",
     "180": "NON_PARTICIPATING",
-    "181": "SUCCESS",
+    "181": "NON_PARTICIPATING",
     "182": "NON_PARTICIPATING",
     "183": "NON_PARTICIPATING",
     "184": "NON_PARTICIPATING",
     "185": "NON_PARTICIPATING",
-    "186": "SUCCESS",
+    "186": "NON_PARTICIPATING",
     "187": "NON_PARTICIPATING",
     "188": "NON_PARTICIPATING",
     "189": "NON_PARTICIPATING",
     "19": "SUCCESS",
-    "190": "SUCCESS",
+    "190": "NON_PARTICIPATING",
     "191": "NON_PARTICIPATING",
     "192": "NON_PARTICIPATING",
     "193": "NON_PARTICIPATING",
@@ -114,24 +114,24 @@
     "197": "NON_PARTICIPATING",
     "198": "NON_PARTICIPATING",
     "199": "NON_PARTICIPATING",
-    "2": "SUCCESS",
+    "2": "NON_PARTICIPATING",
     "20": "SUCCESS",
-    "200": "SUCCESS",
-    "201": "NON_PARTICIPATING",
+    "200": "NON_PARTICIPATING",
+    "201": "SUCCESS",
     "202": "NON_PARTICIPATING",
     "203": "NON_PARTICIPATING",
     "204": "NON_PARTICIPATING",
     "205": "NON_PARTICIPATING",
-    "206": "NON_PARTICIPATING",
+    "206": "SUCCESS",
     "207": "NON_PARTICIPATING",
-    "208": "SUCCESS",
-    "209": "NON_PARTICIPATING",
+    "208": "NON_PARTICIPATING",
+    "209": "SUCCESS",
     "21": "SUCCESS",
-    "210": "NON_PARTICIPATING",
+    "210": "SUCCESS",
     "211": "NON_PARTICIPATING",
     "212": "NON_PARTICIPATING",
     "213": "NON_PARTICIPATING",
-    "214": "NON_PARTICIPATING",
+    "214": "SUCCESS",
     "215": "NON_PARTICIPATING",
     "216": "NON_PARTICIPATING",
     "217": "NON_PARTICIPATING",
@@ -149,7 +149,7 @@
     "228": "SUCCESS",
     "229": "NON_PARTICIPATING",
     "23": "SUCCESS",
-    "230": "SUCCESS",
+    "230": "NON_PARTICIPATING",
     "231": "NON_PARTICIPATING",
     "232": "NON_PARTICIPATING",
     "233": "NON_PARTICIPATING",
@@ -169,7 +169,7 @@
     "246": "NON_PARTICIPATING",
     "247": "NON_PARTICIPATING",
     "248": "NON_PARTICIPATING",
-    "249": "NON_PARTICIPATING",
+    "249": "SUCCESS",
     "25": "SUCCESS",
     "250": "NON_PARTICIPATING",
     "251": "NON_PARTICIPATING",
@@ -181,7 +181,7 @@
     "27": "SUCCESS",
     "28": "SUCCESS",
     "29": "SUCCESS",
-    "3": "NON_PARTICIPATING",
+    "3": "SUCCESS",
     "30": "NON_PARTICIPATING",
     "31": "NON_PARTICIPATING",
     "32": "SUCCESS",
@@ -197,7 +197,7 @@
     "41": "SUCCESS",
     "42": "SUCCESS",
     "43": "SUCCESS",
-    "44": "SUCCESS",
+    "44": "NON_PARTICIPATING",
     "45": "SUCCESS",
     "46": "NON_PARTICIPATING",
     "47": "SUCCESS",
@@ -231,10 +231,10 @@
     "72": "SUCCESS",
     "73": "NON_PARTICIPATING",
     "74": "SUCCESS",
-    "75": "NON_PARTICIPATING",
+    "75": "SUCCESS",
     "76": "SUCCESS",
     "77": "SUCCESS",
-    "78": "NON_PARTICIPATING",
+    "78": "SUCCESS",
     "79": "NON_PARTICIPATING",
     "8": "SUCCESS",
     "80": "SUCCESS",
@@ -247,7 +247,7 @@
     "87": "NON_PARTICIPATING",
     "88": "SUCCESS",
     "89": "SUCCESS",
-    "9": "NON_PARTICIPATING",
+    "9": "SUCCESS",
     "90": "SUCCESS",
     "91": "SUCCESS",
     "92": "SUCCESS",
@@ -268,20 +268,20 @@
     "AutoModelForCausalLM": "distributed/optimized-gpt2-500m--modeling_gpt_optimized.GPTOptim"
   },
   "block_list": [
-    5604697,
-    5604701,
-    5604706,
-    5604710,
-    5604715
+    5606878,
+    5606882,
+    5606887,
+    5606892,
+    5606896
   ],
   "block_size": 1024,
   "bos_token_id": 50256,
   "embd_pdrop": 0.1,
   "eos_token_id": 50256,
   "initializer_range": 0.02,
-  "inner_step": 6,
+  "inner_step": 0,
   "inner_steps": 0,
-  "last_allreduce_block": 5602341,
+  "last_allreduce_block": 5605539,
   "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:a78006948ae14155b3f8ed749cd2943bec4507670cab75c9c0ab451112f5ffc3
+oid sha256:92be508cf8c43a9ca0172681fffd1a9eca6edd4d15790a9d17459e2c9e7b3b54
 size 8081782026

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e3d035a18af093d85963f6e768df907321fed90ae97e76113ce23b9ae488523a
+oid sha256:df791399608bc121428e81d89c881fb0c5bf84b133084b48e6ccb4ed8f64a677
 size 4040701744