New best standard model (full-finetune) at cycle 1 with loss 2.576410 (Trained with Unsloth)

config.json

@@ -1,31 +1,31 @@
 {
+  "_remove_final_layer_norm": false,
+  "activation_dropout": 0.0,
+  "activation_function": "relu",
   "architectures": [
-    "LlamaForCausalLM"
+    "OPTForCausalLM"
   ],
-  "attention_bias": false,
   "attention_dropout": 0.0,
-  "bos_token_id": 0,
-  "eos_token_id": 1,
-  "head_dim": 4,
-  "hidden_act": "silu",
-  "hidden_size": 16,
-  "initializer_range": 0.02,
-  "intermediate_size": 64,
+  "bos_token_id": 2,
+  "do_layer_norm_before": true,
+  "dropout": 0.1,
+  "enable_bias": true,
+  "eos_token_id": 2,
+  "ffn_dim": 3072,
+  "hidden_size": 768,
+  "init_std": 0.02,
+  "layer_norm_elementwise_affine": true,
+  "layerdrop": 0.0,
   "max_position_embeddings": 2048,
-  "mlp_bias": false,
-  "model_type": "llama",
-  "num_attention_heads": 4,
-  "num_hidden_layers": 2,
-  "num_key_value_heads": 4,
-  "pad_token_id": -1,
-  "pretraining_tp": 1,
-  "rms_norm_eps": 1e-06,
-  "rope_scaling": null,
-  "rope_theta": 10000.0,
-  "tie_word_embeddings": false,
+  "model_type": "opt",
+  "num_attention_heads": 12,
+  "num_hidden_layers": 12,
+  "pad_token_id": 1,
+  "prefix": "</s>",
   "torch_dtype": "bfloat16",
   "transformers_version": "4.50.0",
   "unsloth_version": "2025.3.18",
   "use_cache": true,
-  "vocab_size": 32000
+  "vocab_size": 50272,
+  "word_embed_proj_dim": 768
 }

generation_config.json

@@ -1,8 +1,8 @@
 {
   "_from_model_config": true,
-  "bos_token_id": 0,
-  "eos_token_id": 1,
+  "bos_token_id": 2,
+  "eos_token_id": 2,
   "max_length": 2048,
-  "pad_token_id": -1,
+  "pad_token_id": 1,
   "transformers_version": "4.50.0"
 }

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:326f257e14c6b855478f1dbc8bc106c3cb3cd1cef8a1955a520f7a9e336032d5
-size 2066752
+oid sha256:381a5aa2d2444e5debc238689d6bf7fff800cc374493f3161ce17785202a042d
+size 250501160