DeepSeek-R1-Distill-Qwen-14b-chemical-synthesis-PN-adapter

DeepSeek-R1-Distill-Qwen-14B Chemical Synthesis Classifier — LoRA Adapter (Epoch 9)

Model Overview

This repository contains the LoRA adapter for a DeepSeek‑R1‑Distill‑Qwen‑14B model fine‑tuned to classify chemical synthesizability (P = synthesizable, N = unsynthesizable). Training uses QLoRA on an imbalanced P/N dataset; evaluation scores each example with logsumexp(P) − logsumexp(N) at the final token under the SFT‑aligned prompt.

• Task: Binary classification (P vs N)
• Objective: QLoRA with class‑imbalance handling
• Max sequence length (eval): 180 tokens
• Dataset (train): train_llm_pn.jsonl
• Dataset (val): valid_llm_pn.jsonl
• Checkpoint: Epoch 9 (best Balanced Accuracy among evaluated epochs)

The checkpoint includes a chat_template.jinja to ensure prompt formatting matches SFT conditions.

Model Overview

This repository contains the LoRA adapter for the chemical synthesis classifier fine-tuned with focal loss. Training prompts follow the template:

You are a materials science assistant. Given a chemical composition, answer only with 'P' (synthesizable/positive) or 'N' (non-synthesizable/negative)." Correspondingly, each user query was formatted as: "Is the material {composition} likely synthesizable? Answer with P (positive) or N (negative).


Implementation notes for faithful scoring/inference:

- Build inputs via the chat template; drop the final assistant label from dataset inputs and tokenize with an empty assistant turn.
- Use `add_generation_prompt=False` and read logits right after the assistant start (trim the trailing EOS if present).
- Force `attn_implementation="eager"` for stability.

## Validation Metrics (Epoch 4 — Best)

| Metric              | Value      |
| ------------------- | ---------- |
| TPR (P Recall)      | **0.9750** |
| TNR (N Specificity) | 0.9556     |

## How to Load (Transformers + PEFT)

```python
from transformers import AutoModelForCausalLM, AutoTokenizer, BitsAndBytesConfig
from peft import PeftModel

base = "unsloth/DeepSeek-R1-Distill-Qwen-14B-bnb-4bit"  # Base model repo or equivalent local path

bnb = BitsAndBytesConfig(
    load_in_4bit=True,
    bnb_4bit_quant_type="nf4",
    bnb_4bit_use_double_quant=True,
)

tok = AutoTokenizer.from_pretrained(base, use_fast=True)
tok.padding_side = "right"
if tok.pad_token is None:
    tok.pad_token = tok.eos_token

model = AutoModelForCausalLM.from_pretrained(
    base,
    quantization_config=bnb,
    device_map="auto",
    attn_implementation="eager",
)
model = PeftModel.from_pretrained(model, adapter)
model.eval()


## Training Setup (Summary)

- Base model: Unsloth "DeepSeek‑R1‑Distill‑Qwen‑14B‑bnb‑4bit" (4‑bit NF4)
- Fine‑tuning: QLoRA via PEFT
  - Target modules: `q_proj, k_proj, v_proj, o_proj, gate_proj, up_proj, down_proj`
  - LoRA config: `r=32, alpha=32, dropout=0.0`
- Objective: P/U‑only Focal Loss applied to the last P or N token
  - `gamma=2.0`, `alpha_P=7.5`, `alpha_N=1.0`
- Tokenizer: right‑padding; `pad_token = eos_token` if undefined

## Dataset Sources

The training/validation corpus combines multiple public sources and internal curation:

- P/U labelled data from J. Am. Chem. Soc. 2024, 146, 29, 19654-19659 (doi:10.1021/jacs.4c05840)
- High-entropy materials data from Data in Brief 2018, 21, 2664-2678 (doi:10.1016/j.dib.2018.11.111)
- Additional candidates via literature queries and manual screening of high-entropy materials

## VRAM & System Requirements

- GPU VRAM: ≥16 GB recommended (4‑bit base + adapter)
- RAM: ≥16 GB recommended for tokenization and batching
- Libraries: transformers, peft, bitsandbytes (evaluation uses transformers loader)
- Set `attn_implementation="eager"` to avoid SDPA instability

## Limitations & Notes

- The adapter targets chemical synthesizability judgments; generalization outside this domain is not guaranteed.
- For consistent results, use the included `chat_template.jinja` and avoid inserting `<think>` tokens.
- Do not mutate the base model `config.json` (e.g., model_type), which can reinitialize weights and corrupt metrics.