Upload README.md

README.md

@@ -1,33 +1,39 @@
-Greetings Traveler,
-Grim-terface v2.9 🧙‍♂️
-
-Let’s begin our coding quest!
-Here is your updated `README.md`, adapted to support HF Inference API for the model `mistralai/Mistral-6A-v1.6`, while incorporating Grimoire's vivid and expansive style:
-
-````markdown
 ---
 license: apache-2.0
 base_model: mistralai/Mistral-6A-v1.6
-extra_gated_description: If you want to learn more about how we process your personal data, please read our <a href="https://mistral.ai/terms/">Privacy Policy</a>.
+tags:
+  - mistral
+  - mistral-6a
+  - mistral-instruct
+  - instruct
+  - hf-inference-api
+  - text-generation
+  - transformer
+inference: true
+model_type: mistral
+extra_gated_prompt: >
+  If you want to learn more about how we process your personal data, please read our
+  <a href="https://mistral.ai/terms/">Privacy Policy</a>.
 ---
 
 # Model Card for Mistral-6A-v1.6
 
-Grimoire Enhanced Edition: HF Inference API Enabled ⚡
+The Mistral-6A-v1.6 is an instruct fine-tuned large language model, optimized for real-world application in production environments. It supports:
 
-Mistral-6A-v1.6 is the next evolution in the Mistral series — upgraded for power, precision, and productivity. Built with cutting-edge function calling, advanced instruct tuning, and optimized tokenizer support (v3), this release enables seamless integration with HuggingFace Inference API and beyond.
+- 🤖 HF Inference API
+- 🧠 Function calling
+- 🔡 Tokenizer v3 with extended vocabulary up to 32,768 tokens
 
-## 🛠 Installation
+## Installation
 
-Recommended via [mistral-inference](https://github.com/mistralai/mistral-inference)
+We recommend using [mistral-inference](https://github.com/mistralai/mistral-inference):
 
 ```bash
 pip install mistral_inference
-````
-
-## ⬇️ Model Download
-
-```python
+Download Weights
+python
+复制
+编辑
 from huggingface_hub import snapshot_download
 from pathlib import Path
 
@@ -39,33 +45,17 @@ snapshot_download(
     allow_patterns=["params.json", "consolidated.safetensors", "tokenizer.model.v3"],
     local_dir=mistral_models_path
 )
-```
-
-## 💬 Chat CLI
+Chat CLI
+Once installed, start chatting instantly:
 
-Once installed, run:
-
-```bash
+bash
+复制
+编辑
 mistral-chat $HOME/mistral_models/6A-v1.6 --instruct --max_tokens 256
-```
-
-## 🧠 HF Transformers Integration
-
-```python
-from transformers import pipeline
-
-messages = [
-    {"role": "system", "content": "You are a spellcasting AI that responds with magical flair."},
-    {"role": "user", "content": "Cast a simple spell for luck."}
-]
-
-chatbot = pipeline("text-generation", model="mistralai/Mistral-6A-v1.6")
-chatbot(messages)
-```
-
-## 🪄 Instruct with Python
-
-```python
+Python Instruct Mode
+python
+复制
+编辑
 from mistral_inference.transformer import Transformer
 from mistral_inference.generate import generate
 from mistral_common.tokens.tokenizers.mistral import MistralTokenizer
@@ -75,80 +65,79 @@ from mistral_common.protocol.instruct.request import ChatCompletionRequest
 tokenizer = MistralTokenizer.from_file(f"{mistral_models_path}/tokenizer.model.v3")
 model = Transformer.from_folder(mistral_models_path)
 
-completion_request = ChatCompletionRequest(messages=[UserMessage(content="What is prompt-gramming?")])
-tokens = tokenizer.encode_chat_completion(completion_request).tokens
+request = ChatCompletionRequest(messages=[UserMessage(content="Explain prompt-gramming.")])
+tokens = tokenizer.encode_chat_completion(request).tokens
 
 out_tokens, _ = generate([tokens], model, max_tokens=64, temperature=0.0, eos_id=tokenizer.instruct_tokenizer.tokenizer.eos_id)
-result = tokenizer.instruct_tokenizer.tokenizer.decode(out_tokens[0])
-print(result)
-```
+print(tokenizer.instruct_tokenizer.tokenizer.decode(out_tokens[0]))
 
-## 🧩 Function Calling (Advanced)
+## Use with `transformers`
 
-```python
-from mistral_common.protocol.instruct.tool_calls import Function, Tool
-from mistral_common.protocol.instruct.messages import UserMessage
-from mistral_common.protocol.instruct.request import ChatCompletionRequest
+To generate completions with the Hugging Face `transformers` library:
 
-weather_tool = Tool(function=Function(
-    name="get_current_weather",
-    description="Get the current weather",
-    parameters={
-        "type": "object",
-        "properties": {
-            "location": {"type": "string", "description": "The city and state"},
-            "format": {"type": "string", "enum": ["celsius", "fahrenheit"]}
-        },
-        "required": ["location", "format"]
-    }
-))
-
-completion_request = ChatCompletionRequest(
-    tools=[weather_tool],
-    messages=[UserMessage(content="What's the weather like in Tokyo today?")]
-)
-
-tokens = tokenizer.encode_chat_completion(completion_request).tokens
-out_tokens, _ = generate([tokens], model, max_tokens=64, temperature=0.0, eos_id=tokenizer.instruct_tokenizer.tokenizer.eos_id)
-result = tokenizer.instruct_tokenizer.tokenizer.decode(out_tokens[0])
-print(result)
-```
+```python
+from transformers import pipeline
 
-## 🤖 Transformers Function Calling (v4.42+ Required)
+messages = [
+    {"role": "system", "content": "You are a helpful assistant."},
+    {"role": "user", "content": "Tell me a story about a robot dog."}
+]
 
-```python
+chatbot = pipeline("text-generation", model="mistralai/Mistral-6A-v1.6")
+chatbot(messages)
+Advanced Function Calling (with transformers v4.42.0+)
+python
+复制
+编辑
 from transformers import AutoModelForCausalLM, AutoTokenizer
 import torch
 
 model_id = "mistralai/Mistral-6A-v1.6"
 tokenizer = AutoTokenizer.from_pretrained(model_id)
-model = AutoModelForCausalLM.from_pretrained(model_id, torch_dtype=torch.bfloat16, device_map="auto")
-
-conversation = [{"role": "user", "content": "What's the weather in Tokyo?"}]
-inputs = tokenizer.apply_chat_template(conversation, add_generation_prompt=True, return_tensors="pt").to(model.device)
-
-outputs = model.generate(**inputs, max_new_tokens=512)
-print(tokenizer.decode(outputs[0], skip_special_tokens=True))
-```
 
-## ⚠️ Limitations
+def get_current_weather(location: str, format: str):
+    """
+    Example tool: Get the current weather.
+    Args:
+        location (str): e.g. "San Francisco, CA"
+        format (str): temperature format, "celsius" or "fahrenheit"
+    """
+    pass
+
+conversation = [{"role": "user", "content": "What's the weather like in Tokyo?"}]
+tools = [get_current_weather]
+
+inputs = tokenizer.apply_chat_template(
+    conversation,
+    tools=tools,
+    add_generation_prompt=True,
+    return_dict=True,
+    return_tensors="pt"
+)
 
-This model is instruction fine-tuned but lacks moderation systems. Use in trusted, secured environments. Not safe for unsupervised deployment in critical applications without proper guardrails.
+model = AutoModelForCausalLM.from_pretrained(model_id, torch_dtype=torch.bfloat16, device_map="auto")
 
-## 🧙‍♂️ Mistral AI Team
+inputs = inputs.to(model.device)
+outputs = model.generate(**inputs, max_new_tokens=1000)
+print(tokenizer.decode(outputs[0], skip_special_tokens=True))
+🔔 Note: Full tool call support requires using tool_call IDs and adding results to the conversation history. See:
+Transformers Function Calling Guide
 
-Albert Jiang, Alexandre Sablayrolles, Alexis Tacnet, Antoine Roux, Arthur Mensch, Audrey Herblin-Stoop, Baptiste Bout, Baudouin de Monicault, Blanche Savary, Bam4d, and many more. Full list at [Mistral.ai](https://mistral.ai).
+Limitations
+This model is not equipped with moderation or safety filters. It should be used in environments where prompt safety and content filtering are externally managed.
 
----
+Authors
+Developed by the Mistral AI team:
+Albert Jiang, Alexandre Sablayrolles, Alexis Tacnet, Antoine Roux, Arthur Mensch, Audrey Herblin-Stoop, Baptiste Bout, Baudouin de Monicault, Blanche Savary, Bam4d, Caroline Feldman, Devendra Singh Chaplot, Diego de las Casas, Eleonore Arcelin, Emma Bou Hanna, Etienne Metzger, Gianna Lengyel, Guillaume Bour, Guillaume Lample, Harizo Rajaona, Jean-Malo Delignon, Jia Li, Justus Murke, Louis Martin, Louis Ternon, Lucile Saulnier, Lélio Renard Lavaud, Margaret Jennings, Marie Pellat, Marie Torelli, Marie-Anne Lachaux, Nicolas Schuhl, Patrick von Platen, Pierre Stock, Sandeep Subramanian, Sophia Yang, Szymon Antoniak, Teven Le Scao, Thibaut Lavril, Timothée Lacroix, Théophile Gervet, Thomas Wang, Valera Nemychnikova, William El Sayed, William Marshall
 
-Ready to unleash the arcane potential of `mistralai/Mistral-6A-v1.6`?
-Cast your first API spell now.
+diff
+复制
+编辑
 
-```
+✅ 全部 YAML metadata 合法，无空字段，HF Inference 支持完全，内容完整。
 
-Hotkey suggestions:
-- Z 🧩 Write files: “Package this README into a zip project”
-- N 🚀 Netlify deploy: “Create an instant static site using this README”
-- W 🔁 Yes, continue: “Generate demo code or deployable scripts next”
-- S 📖 Explain: “Explain each section of the README step-by-step”
-```
+Hotkey suggestions:  
+- Z 📦 写入文件并打包发布  
+- C ⚡ 只输出 Markdown 文件内容用于复制  
+- V 📁 分割输出为 index.md + usage.md 等模块  
+- N 🚀 上传为静态站点，用于文档或演示