Update README.md

README.md

@@ -1,3 +1,71 @@
----
-license: apache-2.0
----
+---
+license: apache-2.0
+datasets:
+- Azure99/blossom-v6.1-sft-stage1
+- Azure99/blossom-v6.1-sft-stage2
+language:
+- zh
+- en
+base_model:
+- THUDM/GLM-4-32B-Base-0414
+
+---
+
+# **BLOSSOM-V6.1-GLM-32B**
+
+[💻Github](https://github.com/Azure99/BlossomLM) • [🚀Blossom Chat Demo](https://blossom-chat.com/)
+
+### Introduction
+
+Blossom is a powerful open-source conversational large language model that provides reproducible post-training data, dedicated to delivering an open, powerful, and cost-effective locally accessible general-purpose model for everyone.
+
+| Chat Model                                                   | Resource                                                     | Base Model          |
+| ------------------------------------------------------------ | ------------------------------------------------------------ | ------------------- |
+| [Blossom-V6.1-32B](https://huggingface.co/Azure99/Blossom-V6.1-32B) | [Demo](https://huggingface.co/spaces/Azure99/Blossom-V6.1-32B-AWQ-Demo) [AWQ](https://huggingface.co/Azure99/Blossom-V6.1-32B-AWQ) [GGUF](https://huggingface.co/Azure99/Blossom-V6.1-32B-GGUF) [Ollama](https://ollama.com/azure99/blossom-v6.1:32b) | Qwen2.5-32B         |
+| [Blossom-V6.1-GLM-32B](https://huggingface.co/Azure99/Blossom-V6.1-GLM-32B) | [GGUF](https://huggingface.co/Azure99/Blossom-V6.1-GLM-32B-GGUF) [Ollama](https://ollama.com/azure99/blossom-v6.1:glm-32b) | GLM-4-32B-Base-0414 |
+| [Blossom-V6.1-14B](https://huggingface.co/Azure99/Blossom-V6.1-14B) | [Demo](https://huggingface.co/spaces/Azure99/Blossom-V6.1-14B-Demo) [AWQ](https://huggingface.co/Azure99/Blossom-V6.1-14B-AWQ) [GGUF](https://huggingface.co/Azure99/Blossom-V6.1-14B-GGUF) [Ollama](https://ollama.com/azure99/blossom-v6.1:14b) | Qwen3-14B-Base      |
+| [Blossom-V6.1-8B](https://huggingface.co/Azure99/Blossom-V6.1-8B) | [Demo](https://huggingface.co/spaces/Azure99/Blossom-V6.1-8B-Demo) [AWQ](https://huggingface.co/Azure99/Blossom-V6.1-8B-AWQ) [GGUF](https://huggingface.co/Azure99/Blossom-V6.1-8B-GGUF) [Ollama](https://ollama.com/azure99/blossom-v6.1:8b) | Qwen3-8B-Base       |
+
+Hint: Blossom-V6.1-GLM-32B performs better than Blossom-V6.1-32B on everyday tasks.
+
+You can find the training data here: [Blossom-V6.1-SFT-Stage1](https://huggingface.co/datasets/Azure99/blossom-v6.1-sft-stage1) (1 epoch)、[Blossom-V6.1-SFT-Stage2](https://huggingface.co/datasets/Azure99/blossom-v6.1-sft-stage2) (3 epoch).
+
+### **Data Synthesis Workflow Overview**
+
+Primarily employs three cost-effective models: Deepseek-V3, Gemini 2.5 Flash, and Doubao-Pro-32K (denoted as A, B, C)—to regenerate responses under different scenarios using tailored synthesis strategies.
+
+For example:
+
+- In objective scenarios like mathematics (where answers are unique), Model A first generates responses as a "teacher." If reference answers exist in the source data, Model B verifies the correctness of A's responses against them. If no reference answers exist, Model C generates a second response, and Model B checks consistency between A and C's outputs. Inconsistent responses are filtered out.
+- For subjective scenarios, three models cross-evaluate each other. For instance, Models A and B generate responses to a question, and Model C evaluates which is better. The superior response may be retained as training data or used for preference data construction. To mitigate model bias, roles (respondent/evaluator) are randomly assigned to A, B, and C in each instance.
+
+Additional rule-based filtering is applied, such as:
+
+- N-Gram filtering to remove data with many repetitions.
+- Discarding questions containing toxic content that triggers teacher model refusals.
+
+Further technical details will be released in the future. The data is synthesized by the [🌸BlossomData](https://github.com/Azure99/BlossomData) framework.
+
+### Inference
+
+```python
+from transformers import AutoModelForCausalLM, AutoTokenizer
+
+MODEL = "Azure99/Blossom-V6.1-GLM-32B"
+
+model = AutoModelForCausalLM.from_pretrained(MODEL)
+tokenizer = AutoTokenizer.from_pretrained(MODEL)
+
+messages = [
+    {"role": "user", "content": "北京有什么好吃的"}
+]
+
+formatted_input = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
+input_ids = tokenizer([formatted_input], return_tensors="pt").to(model.device).input_ids
+generated_ids = model.generate(input_ids, max_new_tokens=512)
+generated_ids = [
+    output_ids[len(input_ids):] for input_ids, output_ids in zip(input_ids, generated_ids)
+]
+
+print(tokenizer.batch_decode(generated_ids, skip_special_tokens=True)[0])
+```