Training in progress, step 100

config.json

@@ -20,7 +20,7 @@
   "sliding_window": 4096,
   "tie_word_embeddings": false,
   "torch_dtype": "bfloat16",
-  "transformers_version": "4.36.2",
+  "transformers_version": "4.39.0.dev0",
   "use_cache": false,
   "vocab_size": 32000
 }

runs/May01_16-51-03_tongzhang4.math.ust.hk/events.out.tfevents.1714555147.tongzhang4.math.ust.hk.2219749.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:38cbf72f7a0a8d5185ecb04112e5f01ced909e6acd438e09bfec27638bef6b3d
+size 4749

runs/May01_19-38-45_tongzhang4.math.ust.hk/events.out.tfevents.1714563592.tongzhang4.math.ust.hk.2311159.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cb334de88c1fc8c2e3cb7c0bbec10fe296c91f035691501ee6ea7a9f403b6238
+size 6101

runs/May01_19-58-40_tongzhang4.math.ust.hk/events.out.tfevents.1714564785.tongzhang4.math.ust.hk.2321598.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ded5b16a9360e56ce2741c169db588229e3586b9a39eaf696fb74c561efaae63
+size 12912

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c4502f32d21f3a579ae908e694e6ab3a88dfa387ea30b86c3125791941bfe829
-size 5944
+oid sha256:407db9e410c0136a73ffe813ccaed480c6fa357c442b97dd0441891599fa9f2c
+size 6200