From 93d557256fad322be1582abbbd7672e11d7eadfa Mon Sep 17 00:00:00 2001
From: coffee <hanluzhi0429@gmail.com>
Date: Tue, 29 Apr 2025 18:59:54 +0800
Subject: [PATCH] add LLM code

---
 LLM_code/test.py                         |  3 +++
 LLM_code/test_huggingface_train.py       |  0
 LLM_code/test_huggingface_transformer.py | 18 ++++++++++++++++++
 3 files changed, 21 insertions(+)
 create mode 100644 LLM_code/test.py
 create mode 100644 LLM_code/test_huggingface_train.py
 create mode 100644 LLM_code/test_huggingface_transformer.py

diff --git a/LLM_code/test.py b/LLM_code/test.py
new file mode 100644
index 0000000..11acf7e
--- /dev/null
+++ b/LLM_code/test.py
@@ -0,0 +1,3 @@
+from transformers import AutoTokenizer
+tokenizer = AutoTokenizer.from_pretrained("tokenizer_file/hf_tokenizer")  
+print(tokenizer.vocab_size)
diff --git a/LLM_code/test_huggingface_train.py b/LLM_code/test_huggingface_train.py
new file mode 100644
index 0000000..e69de29
diff --git a/LLM_code/test_huggingface_transformer.py b/LLM_code/test_huggingface_transformer.py
new file mode 100644
index 0000000..c4e6ed9
--- /dev/null
+++ b/LLM_code/test_huggingface_transformer.py
@@ -0,0 +1,18 @@
+import transformers
+from transformers import AutoModelForCausalLM, AutoTokenizer
+from transformers import LlamaForCausalLM
+tokenizer = AutoTokenizer.from_pretrained("Qwen/Qwen2-0.5B-Instruct")
+from transformers import LlamaConfig
+config = LlamaConfig(
+    vocab_size=32000,  
+    hidden_size=4096,  
+    num_attention_heads=32,  
+    num_hidden_layers=32, 
+    intermediate_size=11008,  
+)
+
+model = LlamaForCausalLM(config)
+
+model = transformers.Qwen2ForCausalLM(config)
+model_size = sum(t.numel() for t in model.parameters())
+print(f"Model Size: {model_size/1000**2:.1f}M parameters")
\ No newline at end of file