Replace with SQL+Magicoder mixed-training adapter (better SQL generalization + retained code capability, see experiment_log.md)

README.md

@@ -1,102 +1,51 @@
 ---
 base_model: deepseek-ai/deepseek-coder-7b-instruct-v1.5
-library_name: peft
-model_name: sql-adapter
+library_name: transformers
+model_name: deepseek-sql-magicoder-adapter
 tags:
-- base_model:adapter:deepseek-ai/deepseek-coder-7b-instruct-v1.5
-- lora
-- sft
-- text-to-sql
+- generated_from_trainer
 - trl
-license: apache-2.0
-pipeline_tag: text-generation
-datasets:
-- b-mc2/sql-create-context
-- ise-uiuc/Magicoder-OSS-Instruct-75K
+- sft
+licence: license
 ---
 
-# Model Card for sql-adapter
-
-LoRA adapter fine-tuned from [deepseek-ai/deepseek-coder-7b-instruct-v1.5](https://huggingface.co/deepseek-ai/deepseek-coder-7b-instruct-v1.5) for text-to-SQL generation, trained with [TRL](https://github.com/huggingface/trl) SFTTrainer.
-
-## What's new in this version
-
-The first version of this adapter was trained on 100% SQL data (`b-mc2/sql-create-context`) and showed catastrophic forgetting of general code generation ability (HumanEval pass@1 dropped from 60% to 50%, HumanEval+ from 50% to 40%).
-
-This version mixes in code-instruction data during training to mitigate that regression:
-
-| Dataset | Mix ratio | Purpose |
-|---|---|---|
-| [b-mc2/sql-create-context](https://huggingface.co/datasets/b-mc2/sql-create-context) | 30% | Text-to-SQL |
-| [ise-uiuc/Magicoder-OSS-Instruct-75K](https://huggingface.co/datasets/ise-uiuc/Magicoder-OSS-Instruct-75K) | 70% | Preserve general code generation ability |
-
-Mixing was done batch-wise via `datasets.interleave_datasets`. Learning rate was also lowered (2e-4 → 1e-4) and epochs increased (1 → 2) to compensate for the reduced SQL share.
-
-> Note: post-training benchmark numbers for this mixed-data run have not been re-measured yet. The SQL Exact Match / HumanEval numbers below are from the prior 100%-SQL run, kept here for reference; treat this release as a mitigation for known degradation rather than a verified improvement.
-
-## LoRA configuration
-
-```
-r: 16
-lora_alpha: 32
-lora_dropout: 0.05
-target_modules: [q_proj, k_proj, v_proj, o_proj]
-quantization: 4-bit NF4 (QLoRA), bf16 compute
-```
-
-## Prior version results (100% SQL training)
+# Model Card for deepseek-sql-magicoder-adapter
 
-| Metric | Base | SFT | Δ |
-|---|---|---|---|
-| SQL Exact Match (50 samples, sql-create-context) | 4% (2/50) | 78% (39/50) | +74% |
-| HumanEval pass@1 | 60% | 50% | -10% |
-| HumanEval+ (plus) | 50% | 40% | -10% |
+This model is a fine-tuned version of [deepseek-ai/deepseek-coder-7b-instruct-v1.5](https://huggingface.co/deepseek-ai/deepseek-coder-7b-instruct-v1.5).
+It has been trained using [TRL](https://github.com/huggingface/trl).
 
 ## Quick start
 
 ```python
-from transformers import AutoModelForCausalLM, AutoTokenizer
-from peft import PeftModel
-
-base_model_id = "deepseek-ai/deepseek-coder-7b-instruct-v1.5"
-adapter_id = "albertkingdom/deepseek-coder-7b-instruct-sql-create-context-lora"
-
-model = AutoModelForCausalLM.from_pretrained(base_model_id, device_map="auto")
-tokenizer = AutoTokenizer.from_pretrained(base_model_id)
-model = PeftModel.from_pretrained(model, adapter_id)
-
-messages = [{
-    "role": "user",
-    "content": """Given the database schema below, write a SQL query that answers the user's question.
-Only output the SQL query. Do not add any explanation.
-
-### Schema
-CREATE TABLE users (id INT, name VARCHAR(100), email VARCHAR(100))
+from transformers import pipeline
 
-### Question
-Find all users with gmail addresses"""
-}]
-
-inputs = tokenizer.apply_chat_template(messages, return_tensors="pt", add_generation_prompt=True).to(model.device)
-outputs = model.generate(inputs, max_new_tokens=200)
-print(tokenizer.decode(outputs[0], skip_special_tokens=True))
+question = "If you had a time machine, but could only go to the past or the future once and never return, which would you choose and why?"
+generator = pipeline("text-generation", model="albertkingdom/deepseek-sql-magicoder-adapter", device="cuda")
+output = generator([{"role": "user", "content": question}], max_new_tokens=128, return_full_text=False)[0]
+print(output["generated_text"])
 ```
 
 ## Training procedure
 
-Trained with SFT (TRL `SFTTrainer`) on an interleaved SQL + code instruction dataset, 2 epochs, effective batch size 16, bf16, 4-bit QLoRA.
+ 
+
+
+This model was trained with SFT.
 
 ### Framework versions
 
-- PEFT 0.18.0
 - TRL: 0.26.2
 - Transformers: 4.57.3
-- Pytorch: 2.9.1+cu128
+- Pytorch: 2.13.0
 - Datasets: 4.4.2
-- Tokenizers: 0.22.1
+- Tokenizers: 0.22.2
 
 ## Citations
 
+
+
+Cite TRL as:
+    
 ```bibtex
 @misc{vonwerra2022trl,
 	title        = {{TRL: Transformer Reinforcement Learning}},
@@ -106,4 +55,4 @@ Trained with SFT (TRL `SFTTrainer`) on an interleaved SQL + code instruction dat
 	publisher    = {GitHub},
 	howpublished = {\url{https://github.com/huggingface/trl}}
 }
-```
+```

adapter_config.json

@@ -31,8 +31,8 @@
   "target_modules": [
     "q_proj",
     "k_proj",
-    "v_proj",
-    "o_proj"
+    "o_proj",
+    "v_proj"
   ],
   "target_parameters": null,
   "task_type": "CAUSAL_LM",

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:6dc0af55be3a57e2c06c8711404b08f4275720b9c14cf5c0d3d17404efb6d30b
+oid sha256:26b86b81011abc81389fc1caaf1eaf5b4e59a9f4296cfe635a1aec179cf823a8
 size 31489712

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:25941047b55392a17105f7d307e5e32a8ca111a740a6b127f23e5d258067463c
+oid sha256:877b839418ca0ec86b9b662e6b3ffafcd190c512dcf6b00b0828e8f521193da8
 size 6225