Update config.json

config.json

@@ -1,5 +1,5 @@
 {
-  "_name_or_path": "unsloth/qwen2.5-0.5b-instruct-unsloth-bnb-4bit",
+  "_name_or_path": "Qwen/Qwen2.5-0.5B-Instruct",
   "architectures": [
     "Qwen2ForCausalLM"
   ],
@@ -16,15 +16,40 @@
   "num_hidden_layers": 24,
   "num_key_value_heads": 2,
   "pad_token_id": 151654,
+  "quantization_config": {
+    "_load_in_4bit": true,
+    "_load_in_8bit": false,
+    "bnb_4bit_compute_dtype": "bfloat16",
+    "bnb_4bit_quant_storage": "uint8",
+    "bnb_4bit_quant_type": "nf4",
+    "bnb_4bit_use_double_quant": true,
+    "llm_int8_enable_fp32_cpu_offload": false,
+    "llm_int8_has_fp16_weight": false,
+    "llm_int8_skip_modules": [
+      "lm_head",
+      "multi_modal_projector",
+      "merger",
+      "modality_projection",
+      "model.layers.0.self_attn",
+      "model.layers.0.mlp",
+      "model.layers.2.mlp",
+      "model.layers.3.mlp",
+      "model.layers.21.mlp",
+      "model.layers.0.self_attn.q_proj"
+    ],
+    "llm_int8_threshold": 6.0,
+    "load_in_4bit": true,
+    "load_in_8bit": false,
+    "quant_method": "bitsandbytes"
+  },
   "rms_norm_eps": 1e-06,
   "rope_scaling": null,
   "rope_theta": 1000000.0,
   "sliding_window": null,
   "tie_word_embeddings": true,
-  "torch_dtype": "float16",
-  "transformers_version": "4.49.0",
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.49.0.dev0",
   "unsloth_fixed": true,
-  "unsloth_version": "2025.2.15",
   "use_cache": true,
   "use_sliding_window": false,
   "vocab_size": 151936