Training in progress, step 157

config.json

@@ -1,5 +1,4 @@
 {
-  "_flash_attn_2_enabled": true,
   "architectures": [
     "LlamaForCausalLM"
   ],
@@ -24,7 +23,7 @@
   "rope_scaling": null,
   "rope_theta": 10000.0,
   "tie_word_embeddings": true,
-  "torch_dtype": "float32",
+  "torch_dtype": "bfloat16",
   "transformers_version": "4.52.4",
   "use_cache": true,
   "vocab_size": 32018

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:9b5f21005e5dddb39867b15f39d7498996648347517c34f20c4f3de66f686a3b
-size 78437072
+oid sha256:590b2a4708bce14ae5c96b69a18f6b99fc5a29a5045dca788cdf67aaada33fcb
+size 39224640

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:bb82c98b2b446ef4c244934e8fa2b402c4da56ecbd53e95ef252289afcd1be1c
+oid sha256:af5c076571bb7d9ea155891bc4a361ba9ceadcc8894ade1661dfed89779db703
 size 5777