Training in progress, step 500

config.json

@@ -1,7 +1,7 @@
 {
-  "_name_or_path": "/root/trl/results_v2/Qwen_0.5-DPO_5e-7-3ep_0alp_0lam/checkpoint-1000",
+  "_name_or_path": "JayHyeon/Qwen2.5-0.5B-SFT-2e-5-2ep",
   "architectures": [
-    "Qwen2Model"
+    "Qwen2ForCausalLM"
   ],
   "attention_dropout": 0.0,
   "bos_token_id": 151643,
@@ -21,7 +21,7 @@
   "rope_theta": 1000000.0,
   "sliding_window": null,
   "tie_word_embeddings": true,
-  "torch_dtype": "float32",
+  "torch_dtype": "bfloat16",
   "transformers_version": "4.47.0.dev0",
   "use_cache": false,
   "use_mrope": false,

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:170fac463719e3515dce45db7ea731a97c41964b2311307b47d067fd2e14f0a1
-size 1976161736
+oid sha256:4a755cb689f3c33c39520ae8c3ccd65ef610a2659b74c8108fcf377bbed2b811
+size 1260367448

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9a5320ca4174c5d1dd72c7330bd70fcfed9d7698b365441012342df3e75612c7
+size 7736