Upload folder using huggingface_hub

README.md

@@ -1,14 +1,14 @@
 ---
-base_model: deepseek-ai/DeepSeek-R1-Distill-Qwen-7B
+base_model: deepseek-ai/DeepSeek-R1-Distill-Qwen-32B
 tags:
 - mlx
 ---
 
-# mlx-community/DeepSeek-R1-Distill-Qwen-7B-4bit
+# mlx-community/DeepSeek-R1-Distill-Qwen-32B-4bit
 
-The Model [mlx-community/DeepSeek-R1-Distill-Qwen-7B-4bit](https://huggingface.co/mlx-community/DeepSeek-R1-Distill-Qwen-7B-4bit) was
-converted to MLX format from [deepseek-ai/DeepSeek-R1-Distill-Qwen-7B](https://huggingface.co/deepseek-ai/DeepSeek-R1-Distill-Qwen-7B)
-using mlx-lm version **0.21.1**.
+The Model [mlx-community/DeepSeek-R1-Distill-Qwen-32B-4bit](https://huggingface.co/mlx-community/DeepSeek-R1-Distill-Qwen-32B-4bit) was
+converted to MLX format from [deepseek-ai/DeepSeek-R1-Distill-Qwen-32B](https://huggingface.co/deepseek-ai/DeepSeek-R1-Distill-Qwen-32B)
+using mlx-lm version **0.20.2**.
 
 ## Use with mlx
 
@@ -19,14 +19,14 @@ pip install mlx-lm
 ```python
 from mlx_lm import load, generate
 
-model, tokenizer = load("mlx-community/DeepSeek-R1-Distill-Qwen-7B-4bit")
+model, tokenizer = load("mlx-community/DeepSeek-R1-Distill-Qwen-32B-4bit")
 
-prompt = "hello"
+prompt="hello"
 
-if tokenizer.chat_template is not None:
+if hasattr(tokenizer, "apply_chat_template") and tokenizer.chat_template is not None:
     messages = [{"role": "user", "content": prompt}]
     prompt = tokenizer.apply_chat_template(
-        messages, add_generation_prompt=True
+        messages, tokenize=False, add_generation_prompt=True
     )
 
 response = generate(model, tokenizer, prompt=prompt, verbose=True)

config.json

@@ -6,15 +6,15 @@
     "bos_token_id": 151643,
     "eos_token_id": 151643,
     "hidden_act": "silu",
-    "hidden_size": 3584,
+    "hidden_size": 5120,
     "initializer_range": 0.02,
-    "intermediate_size": 18944,
+    "intermediate_size": 27648,
     "max_position_embeddings": 131072,
-    "max_window_layers": 28,
+    "max_window_layers": 64,
     "model_type": "qwen2",
-    "num_attention_heads": 28,
-    "num_hidden_layers": 28,
-    "num_key_value_heads": 4,
+    "num_attention_heads": 40,
+    "num_hidden_layers": 64,
+    "num_key_value_heads": 8,
     "quantization": {
         "group_size": 64,
         "bits": 4
@@ -23,14 +23,13 @@
         "group_size": 64,
         "bits": 4
     },
-    "rms_norm_eps": 1e-06,
-    "rope_theta": 10000,
-    "sliding_window": 4096,
+    "rms_norm_eps": 1e-05,
+    "rope_theta": 1000000.0,
+    "sliding_window": 131072,
     "tie_word_embeddings": false,
     "torch_dtype": "bfloat16",
-    "transformers_version": "4.44.0",
+    "transformers_version": "4.43.1",
     "use_cache": true,
-    "use_mrope": false,
     "use_sliding_window": false,
     "vocab_size": 152064
 }

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:75ed9ed7ad66720041dd186dd0d8ce8a2dd8f487ea3f45a4b8fcaa7bbbf1f64f
-size 4284346255
+oid sha256:3985d69b00a8f8af6f040fd4e05ab95fc6c0a51e739f37864c29756d5b6e9e2b
+size 4284346187