Update README.md

README.md

@@ -4,7 +4,7 @@ language:
 - zh
 - en
 tags:
-- qwen
+- gemma
 - sales
 - unsloth
 - lora
@@ -15,7 +15,7 @@ tags:
 
 **Model ID:** aifeifei798/QiMing-Gemma-3-4b
 
-**Base Model:** Qwen3-4B (Fine-tuned on a consumer-grade GPU by injecting structural logic)
+**Base Model:** google/gemma-3-4b-it-qat-q4_0-unquantized (Fine-tuned on a consumer-grade GPU by injecting structural logic)
 
 <br>
 
@@ -57,58 +57,6 @@ It is this internal "synergistic operation" that fills Qiming's responses with *
 
 ---
 
-## 🚀 How to Use
-
-```python
-from transformers import AutoModelForCausalLM, AutoTokenizer
-
-model_name = "aifeifei798/QiMing-Gemma-3-4b"
-
-# load the tokenizer and the model
-tokenizer = AutoTokenizer.from_pretrained(model_name)
-model = AutoModelForCausalLM.from_pretrained(
-    model_name,
-    torch_dtype="auto",
-    device_map="auto"
-)
-
-# prepare the model input
-prompt = "My son is in the fifth grade. He's very smart, but he's lost interest in all of his school subjects, and his grades have been slipping. Recently, he's become obsessed with a very complex sandbox game where he builds all sorts of intricate machines. I'm very anxious. On one hand, I'm worried about his academic performance; on the other, I have a gut feeling that I shouldn't crush his creativity. What on earth should I do?"
-messages = [
-    {"role": "system", "content": "You are a helpful assistant."},
-    {"role": "user", "content": prompt}
-]
-text = tokenizer.apply_chat_template(
-    messages,
-    tokenize=False,
-    add_generation_prompt=True,
-)
-model_inputs = tokenizer([text], return_tensors="pt").to(model.device)
-
-# conduct text completion
-generated_ids = model.generate(
-    **model_inputs,
-    max_new_tokens=32768
-)
-output_ids = generated_ids[0][len(model_inputs.input_ids[0]):].tolist() 
-
-# parsing thinking content
-try:
-    # rindex finding 151668 (</think>)
-    index = len(output_ids) - output_ids[::-1].index(151668)
-except ValueError:
-    index = 0
-
-thinking_content = tokenizer.decode(output_ids[:index], skip_special_tokens=True).strip("\n")
-content = tokenizer.decode(output_ids[index:], skip_special_tokens=True).strip("\n")
-
-print("thinking content:", thinking_content) # no opening <think> tag
-print("content:", content)
-
-```
-
----
-
 ## Showcase: An S-Class Maiden Voyage
 
 To validate Qiming's capabilities, we presented it with an exceptionally complex, real-world dilemma that blends education, psychology, and family dynamics.
@@ -224,60 +172,6 @@ Its methodology, training data, and origin story are open-source, in the hope of
 
 ---
 
-## 🚀 使用方法 (How to Use)
-
-本模型是使用 `unsloth` 进行LoRA微调的。为了获得最佳效果，建议使用 `unsloth` 加载模型。
-
-```python
-from transformers import AutoModelForCausalLM, AutoTokenizer
-
-model_name = "aifeifei798/QiMing-Gemma-3-4b"
-
-# load the tokenizer and the model
-tokenizer = AutoTokenizer.from_pretrained(model_name)
-model = AutoModelForCausalLM.from_pretrained(
-    model_name,
-    torch_dtype="auto",
-    device_map="auto"
-)
-
-# prepare the model input
-prompt = "我的孩子今年上五年级，他非常聪明，但对学校的所有科目都失去了兴趣，成绩一直在下滑。最近他迷上了玩一款很复杂的沙盒游戏，在里面建造各种精巧的机器。我非常焦虑，我一方面担心他的学业，另一方面又隐约觉得不该扼杀他的创造力。我到底该怎么办？"
-messages = [
-    {"role": "system", "content": "You are a helpful assistant."},
-    {"role": "user", "content": prompt}
-]
-text = tokenizer.apply_chat_template(
-    messages,
-    tokenize=False,
-    add_generation_prompt=True,
-)
-model_inputs = tokenizer([text], return_tensors="pt").to(model.device)
-
-# conduct text completion
-generated_ids = model.generate(
-    **model_inputs,
-    max_new_tokens=32768
-)
-output_ids = generated_ids[0][len(model_inputs.input_ids[0]):].tolist() 
-
-# parsing thinking content
-try:
-    # rindex finding 151668 (</think>)
-    index = len(output_ids) - output_ids[::-1].index(151668)
-except ValueError:
-    index = 0
-
-thinking_content = tokenizer.decode(output_ids[:index], skip_special_tokens=True).strip("\n")
-content = tokenizer.decode(output_ids[index:], skip_special_tokens=True).strip("\n")
-
-print("thinking content:", thinking_content) # no opening <think> tag
-print("content:", content)
-
-```
-
----
-
 ## 案例展示：一次S级的首航任务
 
 为了验证“启明”的能力，我们向它提出了一个极其复杂的、融合了教育、心理和家庭关系的真实困境。