Outer Step 9. Inner Step 4. Batch Size 128

config.json

@@ -267,13 +267,24 @@
     "AutoConfig": "distributed/optimized-gpt2-500m--configuration_gpt_optimized.GPTOptimConfig",
     "AutoModelForCausalLM": "distributed/optimized-gpt2-500m--modeling_gpt_optimized.GPTOptim"
   },
-  "block_list": [],
+  "block_list": [
+    5314921,
+    5314924,
+    5314926,
+    5314929,
+    5314931,
+    5314934,
+    5314936,
+    5314938,
+    5314940,
+    5314943
+  ],
   "block_size": 1024,
   "bos_token_id": 50256,
   "embd_pdrop": 0.1,
   "eos_token_id": 50256,
   "initializer_range": 0.02,
-  "inner_step": 0,
+  "inner_step": 4,
   "inner_steps": 0,
   "last_allreduce_block": 5297756,
   "layer_norm_epsilon": 1e-05,

inner_optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d241a16916cb5a41ac6f7276ebeabdd3b67a2aff7c39511e0c4239428992fb55
-size 2944
+oid sha256:0e2ce3c920b8fb5f6b8a9e28c70321b7e64fb6434687027ca138ef38b413c3f5
+size 8081782026

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:fb2d6803d7f8fa6434db62fea3b146b1f4ff23b347ced174e318c748a5efaa80
+oid sha256:bc13cdadc10a901688dff9867468d2ef7bc81ad8f6d254327f8a70113421d9bc
 size 4040701744