viggo_finetuning_analysis_batch4_wd0_illconditioned

config.json

@@ -24,7 +24,7 @@
   "rotary_emb_base": 10000,
   "rotary_pct": 0.25,
   "tie_word_embeddings": false,
-  "torch_dtype": "bfloat16",
+  "torch_dtype": "float32",
   "transformers_version": "4.47.1",
   "use_cache": true,
   "use_parallel_residual": true,

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:7764d93555ddf2f29c3ede1456e9eeb311240f59f0aa9fc5aea784b7b1c43cfa
-size 28144000
+oid sha256:d7a21fa9b1b68df79c843a2d1a87448aa82fde22112319b65e55573f38fb36d6
+size 56279344

runs/Jan31_21-16-12_f3debe15d28e/events.out.tfevents.1738358174.f3debe15d28e.1002.7

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:66a343d6a8336104c5dfe77685fa523b5f648e23a6b767509b8b7f6ae09736a6
+size 12908

runs/Jan31_21-16-45_f3debe15d28e/events.out.tfevents.1738358208.f3debe15d28e.1002.8

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6811c067c4aa5e0a53d7cc948b8c767299105679e053541dbfbdda539a35beba
+size 19702

runs/Jan31_21-20-31_f3debe15d28e/events.out.tfevents.1738358433.f3debe15d28e.1002.9

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:78c1405c909a31c7b800c7ebc82ee3d2fa48e47851b265b6c28911b8f3d47217
+size 19742

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:b35c2d9967b2618a45a8f55e2e316cb7e75ed999047b8dc109e137d366a0b1a4
-size 5624
+oid sha256:3c12064d6796a02227c04c13bfcf80437f11a04b439628600ec60d8a8f886d30
+size 5560