Training in progress, step 500, checkpoint

checkpoint-500/config.json

@@ -16,7 +16,7 @@
   "intermediate_size": 3072,
   "layer_norm_eps": 1e-12,
   "max_blocks": 200,
-  "max_position_embeddings": 1801,
+  "max_position_embeddings": 1805,
   "max_relative_position_embeddings": 8,
   "model_type": "longelm",
   "node_size": 1,

checkpoint-500/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:eb433f7e50c675d7f9340210f91030f2b285798933fe184ac4451d7730542a47
-size 1010407418
+oid sha256:e8c93e7ded0c0f6ae327bad94a950037cb6a7278c0784f58d37e4b31eba11d7c
+size 1010431994

checkpoint-500/pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:64ca4e0c617305624bbf4302c1cfe94b9f4e984343dd4eb36432c69d714aa908
-size 505188394
+oid sha256:46fbcc81e33d17b3fe1f9a1f62150621df48633d18586979ced58adae1d3c13c
+size 505200682

checkpoint-500/trainer_state.json

@@ -9,11 +9,11 @@
   "is_world_process_zero": true,
   "log_history": [
     {
-      "ep_loss": 4.7343,
+      "ep_loss": 4.4238,
       "epoch": 0.04,
       "learning_rate": 2.5e-05,
-      "loss": 11.6218,
-      "mlm_loss": 6.8875,
+      "loss": 11.1467,
+      "mlm_loss": 6.7229,
       "step": 500
     }
   ],

checkpoint-500/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:cee56d4f38339eac5ecf098381d5222b3783455863547a17ac0df67c8caa7d15
+oid sha256:b26e03be70ebe5d40a81a81be94421c0578a9d8742b065bc029501df48632957
 size 4664