Upload folder using huggingface_hub

.gitattributes

@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+training_progression.png filter=lfs diff=lfs merge=lfs -text

README.md

@@ -2,45 +2,88 @@
 license: apache-2.0
 base_model: LiquidAI/LFM2.5-1.2B-Instruct
 tags:
-- text-to-sql
-- sql
-- fine-tuned
+  - text2sql
+  - sql
+  - fine-tuned
+  - lora
+  - pytorch
+datasets:
+  - synthetic
 language:
-- en
+  - en
 pipeline_tag: text-generation
-library_name: transformers
 ---
 
-# LFM2.5-1.2B-Text2SQL
+# LFM2.5-1.2B-Text2SQL (PyTorch)
 
-Fine-tuned [LiquidAI/LFM2.5-1.2B-Instruct](https://huggingface.co/LiquidAI/LFM2.5-1.2B-Instruct) for text-to-SQL.
+A fine-tuned version of [LiquidAI/LFM2.5-1.2B-Instruct](https://huggingface.co/LiquidAI/LFM2.5-1.2B-Instruct) for Text-to-SQL generation.
 
-## Performance (vs Teacher: DeepSeek V3)
+## Model Description
 
-| Metric | Base | **Finetuned** | Teacher |
-|--------|------|---------------|---------|
-| Exact Match | 48% | **66%** | 60% |
-| LLM-as-Judge | 75% | **87%** | 90% |
-| ROUGE-L | 0.830 | **0.931** | 0.917 |
-| BLEU | 0.695 | **0.870** | 0.852 |
+This model was fine-tuned on 2000 synthetic Text-to-SQL examples generated using a teacher model (DeepSeek V3).
+The fine-tuning was performed using LoRA adapters with MLX on Apple Silicon, then fused into the base model.
 
-## Usage with vLLM
+### Training Details
+
+- **Base Model**: LiquidAI/LFM2.5-1.2B-Instruct
+- **Training Data**: 2000 synthetic examples
+- **Training Method**: LoRA fine-tuning (FP16)
+- **Iterations**: 5400
+- **Hardware**: Apple Silicon (MLX)
+
+## Performance
+
+### Model Comparison
+
+![Model Comparison](model_comparison.png)
+
+| Metric | Teacher (DeepSeek V3) | Base Model | Fine-tuned |
+|--------|----------------------|------------|------------|
+| Exact Match | 60% | 48% | **72%** |
+| LLM-as-Judge | 90% | 75% | 87% |
+| ROUGE-L | 92% | 83% | **94%** |
+| BLEU | 85% | 70% | **89%** |
+| Semantic Similarity | 96% | 93% | **97%** |
+
+### Training Progression
+
+![Training Progression](training_progression.png)
+
+The model shows consistent improvement across all checkpoints with no signs of overfitting.
+
+## Usage
+
+### PyTorch / Transformers
 
 ```python
-from vllm import LLM, SamplingParams
-
-llm = LLM(model="hybridaione/LFM2.5-1.2B-Text2SQL")
-prompt = '''<|im_start|>system
-You are an expert SQL writer.<|im_end|}
-<|im_start|>user
-Schema:
-CREATE TABLE users (id INTEGER, name TEXT);
-
-Question: Count all users<|im_end|>
-<|im_start|>assistant
-'''
-output = llm.generate([prompt], SamplingParams(temperature=0, max_tokens=256))
+from transformers import AutoModelForCausalLM, AutoTokenizer
+import torch
+
+model = AutoModelForCausalLM.from_pretrained(
+    "furukama/LFM2.5-1.2B-Text2SQL",
+    trust_remote_code=True,
+    torch_dtype=torch.bfloat16,
+    device_map="auto"
+)
+tokenizer = AutoTokenizer.from_pretrained("furukama/LFM2.5-1.2B-Text2SQL", trust_remote_code=True)
+
+# Example query
+prompt = '''CREATE TABLE employees (id INT, name VARCHAR, salary DECIMAL);
+
+Question: What are the names of employees earning more than 50000?'''
+
+messages = [{"role": "user", "content": prompt}]
+inputs = tokenizer.apply_chat_template(messages, return_tensors="pt").to(model.device)
+outputs = model.generate(inputs, max_new_tokens=256)
+print(tokenizer.decode(outputs[0], skip_special_tokens=True))
 ```
 
-## Other Formats
-- **MLX (Apple Silicon)**: [hybridaione/LFM2.5-1.2B-Text2SQL-MLX](https://huggingface.co/hybridaione/LFM2.5-1.2B-Text2SQL-MLX)
+## Limitations
+
+- Trained on synthetic data for a specific database schema
+- Best suited for similar SQL query patterns seen during training
+- May not generalize well to very different database schemas
+
+## License
+
+This model is released under the Apache 2.0 license, following the base model's license.

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e7dd4935411cecb0abf5ac7c7ff34ecdf462cf6d39d77d7454c55b4385531215
+oid sha256:99720bd0525951742d0ce4d753a23bdeaed5fdfacc8c47a49254348995e12e91
 size 2340697904