Trained on 2025-02-27/16-29-15 https://wandb.ai/austinleedavis/train/runs/cb3q0p93

config.json

@@ -1,5 +1,5 @@
 {
-  "_name_or_path": "outputs/2025-02-26/13-23-58/checkpoint/../hf_model",
+  "_name_or_path": "outputs/checkpoint/../hf_model",
   "activation_function": "gelu_new",
   "architectures": [
     "GPT2LMHeadModel"
@@ -10,6 +10,7 @@
   "eos_token_id": 2,
   "initializer_range": 0.02,
   "layer_norm_epsilon": 1e-05,
+  "loss_type": "ForCausalLM",
   "model_type": "gpt2",
   "n_ctx": 1024,
   "n_embd": 768,

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:94f48cb45c303c664768a40bf0831abeb1f93eb1dd4cc57c96a0976d83619004
+oid sha256:33e61e0cd2b46e0ac3db0bc3a4b6fdab3ac300f3cfbf51bae7fc26538b735797
 size 343605888