Update README.md

README.md

@@ -1,60 +1,189 @@
 ---
-library_name: transformers
-license: other
-base_model: QizhiPei/BioMatrix-4B-Base
+license: apache-2.0
+language:
+- en
 tags:
-- llama-factory
-- full
-- generated_from_trainer
-model-index:
-- name: sft_pk_checkpoint_cpt_4b_biom_cpt_ds_32gpus_z0_merge_all_v1_ml2048_eps5_seed42
-  results: []
+- biology
+- chemistry
+- molecule
+- protein
+- multimodal
+- foundation-model
+- drug-discovery
+- protein-design
+pipeline_tag: text-generation
+base_model: QizhiPei/BioMatrix-4B-Base
+library_name: transformers
 ---
 
-<!-- This model card has been generated automatically according to the information the Trainer had access to. You
-should probably proofread and complete it, then remove this comment. -->
+# BioMatrix-4B-SFT
+
+**BioMatrix** is a multimodal biological foundation model that natively integrates **1D sequences**, **3D structures**, and **natural language** for both **molecules** and **proteins** within a single decoder-only architecture.
+
+This is the **4B-parameter SFT (Supervised Fine-Tuned)** variant, instruction-tuned across 80 downstream biological tasks spanning 6 categories.
+
+- 📄 **Paper**: [BioMatrix: Towards a Comprehensive Biological Foundation Model Spanning the Modality Matrix of Sequences, Structures, and Language](https://arxiv.org/abs/xxxx.xxxxx)
+- 💻 **Code**: [https://github.com/QizhiPei/biomatrix](https://github.com/QizhiPei/biomatrix)
+- 🤗 **Model & Data Collection**: [https://huggingface.co/collections/QizhiPei/biomatrix](https://huggingface.co/collections/QizhiPei/biomatrix)
+
+## Model Overview
+
+BioMatrix closes the gap between native multimodality and broad entity coverage in biological foundation models. Unlike adapter-based approaches that bolt external encoders onto a language model, or prior native-tokenization models confined to a single entity type, BioMatrix maps **all modalities into a shared discrete token space** via a unified tokenization scheme:
+
+- **Molecular 1D sequences** (both SMILES and SELFIES notations)
+- **Molecular 3D structures** (via MolStrucTok with branch-decoupled decoder)
+- **Protein 1D sequences** (residue-level tokens)
+- **Protein 3D structures** (via GCP-VQVAE backbone tokenizer)
+- **Natural language** (inherited from Qwen3 tokenizer)
+
+All modalities are consumed and produced uniformly under a **single next-token prediction objective**—without external encoders, projection adapters, or modality-specific output heads.
+
+| Model | Molecule 1D | Molecule 3D | Protein 1D | Protein 3D | Natural Language |
+|-------|:-----------:|:-----------:|:----------:|:----------:|:----------------:|
+| ESM3 | ✗ | ✗ | ✓ | ✓ | ✓ |
+| 3D-MoLM | ✓ | ✓ | ✗ | ✗ | ✓ |
+| AlphaFold3 | ✓ | ✓ | ✓ | ✓ | ✗ |
+| BioT5/BioT5+ | ✓ | ✗ | ✓ | ✗ | ✓ |
+| BioMedGPT | ✓ | ✗ | ✓ | ✗ | ✓ |
+| NatureLM | ✓ | ✗ | ✓ | ✗ | ✓ |
+| SciReasoner | ✓ | ✗ | ✓ | ✗ | ✓ |
+| **BioMatrix** | **✓** | **✓** | **✓** | **✓** | **✓** |
+
+## Model Details
+
+- **Base Architecture**: Qwen3-4B-Base
+- **Parameters**: 4B
+- **Training Stages**:
+  - **Continual Pretraining** on 304.4B tokens (general/scientific text, molecular & protein 1D/3D data, cross-modal interleaved corpora)
+  - **Instruction Tuning** on a comprehensive suite of 80 downstream tasks across 6 categories
+- **Context Length**: 8,192 tokens
+- **Tokenizer**: Extended Qwen3 vocabulary with:
+  - 11,294 joint molecular 3D tokens (composed from SELFIES atom × MolStrucTok codes)
+  - 4,096 protein 3D tokens (GCP-VQVAE codebook)
+  - 26 protein 1D tokens (amino acids + non-standard/unknown)
+  - SELFIES atom tokens and modality-specific control tokens
+
+## Pretraining Corpus (304.4B tokens)
+
+| Category | Tokens | Sources |
+|----------|--------|---------|
+| **Text** | 105.3B | FineWeb-Edu, FineFineWeb (biology/chemistry/medical/health), PubMed Full Articles |
+| **Molecule** | 73.7B | PubChem, PCQM4Mv2, PubChemQC, MolTextNet |
+| **Protein** | 77.4B | UniRef50, RCSB PDB, Swiss-Prot, TrEMBL, AlphaFold DB |
+| **Cross-entity** | 48.0B | Interleaved text (PubMed, bioRxiv, S2ORC, USPTO), Molecule–protein (BindingDB, STITCH, jglaser, CrossDocked), Protein–protein (AlphaSeq, PPIRef) |
+
+## Performance Highlights
+
+BioMatrix achieves **state-of-the-art or competitive performance on 77 out of 80 tasks**. Selected highlights for the 4B-SFT variant:
+
+### Molecular Tasks
+- **Unconditional 1D Generation** (GuacaMol): 0.998 validity, 1.000 uniqueness, 0.986 novelty
+- **Name Conversion (I2S EM)**: 92.83% (vs. SciReasoner-8B: 84.40%)
+- **Text-Based Molecule Generation (EM)**: 65.07% (vs. SciReasoner-8B: 48.00%)
+- **MoleculeQA Total Accuracy**: 73.78% (vs. prior best MolCA-1.3B: 64.79%)
+- **Property-Conditioned 3D Generation**: ~3-4× error reduction on QM9 electronic-structure targets
+
+### Protein Tasks
+- **Fold Type Prediction (Family level)**: 87.25% accuracy
+- **Annotation Prediction (UniProtSeq Keywords F1)**: 91.26%
+- **Inverse Folding AAR**: 75.50% (vs. DPLM-2-3B: 61.67%)
+- **Sequence–Structure Co-generation**: scTM = 0.965, scRMSD = 2.80
+- **Unconditional Backbone Generation**: scTM = 0.963 (joint frontier with RFDiffusion)
+
+### Interaction Tasks
+- **BindingDB Affinity (RMSE)**: 1.030 (new SOTA, surpasses prior literature SOTA of 1.340)
+- **PDBBindv2020 3D Affinity**: RMSE = 1.260, Pearson = 0.737, MAE = 0.972
+
+## Quick Start
+
+```python
+from transformers import AutoModelForCausalLM, AutoTokenizer
+
+model_name = "QizhiPei/BioMatrix-4B-SFT"
+tokenizer = AutoTokenizer.from_pretrained(model_name, trust_remote_code=True)
+model = AutoModelForCausalLM.from_pretrained(
+    model_name,
+    torch_dtype="auto",
+    device_map="auto",
+    trust_remote_code=True
+)
+
+# Example: Molecule captioning with SELFIES input
+instruction = "I need a brief explanation of the molecule denoted in this SELFIES notation. <|mol_sfi_start|>[Te]<|mol_sfi_end|>"
+
+messages = [
+    {"role": "user", "content": instruction}
+]
+prompt = tokenizer.apply_chat_template(
+    messages,
+    tokenize=False,
+    add_generation_prompt=True
+)
+
+inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
+outputs = model.generate(**inputs, max_new_tokens=512, do_sample=False)
+response = tokenizer.decode(outputs[0][inputs.input_ids.shape[1]:], skip_special_tokens=False)
+print(response)
+```
+
+## Modality Wrapping
+
+When constructing prompts, biomolecular content must be wrapped with the corresponding control tokens:
+
+| Modality | Wrapping Example |
+|----------|------------------|
+| Molecule SMILES | `<\|mol_smi_start\|>CC#CC#N<\|mol_smi_end\|>` |
+| Molecule SELFIES | `<\|mol_sfi_start\|>[C][#C][C][#N]<\|mol_sfi_end\|>` |
+| Molecule 3D | `<\|mol_3d_start\|>[H 3][C 0][#C 6]...<\|mol_3d_end\|>` |
+| Protein 1D | `<\|prot_aa_start\|><A M><A R><A A>...<\|prot_aa_end\|>` |
+| Protein 3D | `<\|prot_3d_start\|><S 4012><S 153><S 2091>...<\|prot_3d_end\|>` |
+
+Natural language text is left unwrapped and serves as the default carrier modality.
+
+## Supported Tasks
+
+BioMatrix-4B-SFT was instruction-tuned across the following task categories:
+
+**Molecule (1D)**: unconditional generation, name conversion, property prediction, captioning, text-based generation, forward/retrosynthesis, editing, optimization, customized generation, question answering
 
-# sft_pk_checkpoint_cpt_4b_biom_cpt_ds_32gpus_z0_merge_all_v1_ml2048_eps5_seed42
+**Molecule (3D)**: unconditional generation, property-conditioned generation
 
-This model is a fine-tuned version of [QizhiPei/BioMatrix-4B-Base](https://huggingface.co/QizhiPei/BioMatrix-4B-Base) on the guacamol_smi, the guacamol_sfi, the moses_smi, the moses_sfi, the pdbbind_v2020_mol_first_train, the pdbbind_v2020_mol_first_train, the pdbbind_v2020_mol_first_train, the pdbbind_v2020_mol_first_train, the knowmol_smi_history, the knowmol_smi_nohistory, the knowmol_sfi_history, the knowmol_sfi_nohistory, the mol_smi_all_dy, the mol_sfi_all_dy, the pro_all_zm_v1, the megascience, the yeast_peer, the yeast_peer, the human_peer, the human_peer, the ppi_affinity_peer, the ppi_affinity_peer, the pdbbind_peer, the pdbbind_peer, the bindingdb_peer, the bindingdb_peer, the qm9_2014_uncond, the qm9_2014_uncond, the qm9_2014_uncond, the qm9_2014_uncond, the qm9_2014_condx10, the qm9_2014_condx10, the qm9_2014_condx10, the pfud_3d_1d, the pfud_3d_1d, the pfud_3d_1d, the pfud_3d_1d, the dplm_1d_to_3d, the dplm_1d_to_3d, the dplm_1d_to_3d, the dplm_1d_to_3d, the dplm_1d_to_3d, the dplm_1d_to_3d, the dplm_1d_to_3d, the dplm_1d_to_3d, the dplm_3d_to_1d, the dplm_3d_to_1d, the dplm_3d_to_1d, the dplm_3d_to_1d, the dplm_3d_to_1d, the dplm_3d_to_1d, the dplm_3d_to_1d, the dplm_3d_to_1d, the dplm_1d_3d, the dplm_1d_3d, the dplm_1d_3d, the dplm_1d_3d, the dplm_1d_3d, the dplm_1d_3d, the dplm_1d_3d, the dplm_1d_3d, the dplm_3d_1d, the dplm_3d_1d, the dplm_3d_1d, the dplm_3d_1d, the dplm_3d_1d, the dplm_3d_1d, the dplm_3d_1d and the dplm_3d_1d datasets.
+**Protein (1D)**: sequence understanding, annotation prediction, knowledge mining, text-based design, unconditional generation
 
-## Model description
+**Protein (3D)**: structure understanding, folding, inverse folding, sequence-structure co-generation, unconditional backbone generation
 
-More information needed
+**Interaction**: molecule-protein binding affinity (1D & 3D), protein-protein interaction
 
-## Intended uses & limitations
+> **Note on task-group variants**: As detailed in the paper, the released SFT model is trained on the union of all sub-task corpora with mild oversampling for small-data tasks. For best performance on specific benchmarks, please refer to the paper's task-group-specific variants.
 
-More information needed
+## SMILES vs. SELFIES
 
-## Training and evaluation data
+BioMatrix supports both notations as parallel 1D molecular representations. Empirically:
 
-More information needed
+- **SELFIES** excels on tasks requiring validity-by-construction (unconditional generation, property optimization)
+- **SMILES** excels on tasks requiring surface-level structural anchoring (customized generation with atom/bond/functional-group constraints, forward synthesis, retrosynthesis)
 
-## Training procedure
+See Section 9.2 of the paper for detailed analysis.
 
-### Training hyperparameters
+## Limitations
 
-The following hyperparameters were used during training:
-- learning_rate: 5e-05
-- train_batch_size: 32
-- eval_batch_size: 8
-- seed: 42
-- distributed_type: multi-GPU
-- num_devices: 8
-- total_train_batch_size: 256
-- total_eval_batch_size: 64
-- optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
-- lr_scheduler_type: cosine_with_min_lr
-- lr_scheduler_warmup_ratio: 0.1
-- num_epochs: 5.0
+- Molecular and protein 3D structures are tokenized in **disjoint geometric reference frames**, so the model cannot natively represent biomolecular complexes (e.g., docking poses).
+- Heavy domain specialization may erode some general-purpose language capabilities of the underlying Qwen3 backbone.
+- Coverage is limited to **small molecules and proteins**; nucleic acids, carbohydrates, and lipids are not currently supported.
+- Fine-grained 3D geometry (e.g., bond lengths) shows residual quantization error from finite codebooks; a lightweight post-hoc force-field refinement (e.g., MMFF) closes most of this gap.
 
-### Training results
+## Citation
 
+If you find BioMatrix useful, please cite:
 
+```bibtex
+@article{pei2026biomatrix,
+  title={BioMatrix: Towards a Comprehensive Biological Foundation Model Spanning the Modality Matrix of Sequences, Structures, and Language},
+  author={Pei, Qizhi and Zhou, Zhimeng and Duan, Yi and Zhao, Yiyang and He, Liang and Hsieh, Chang-Yu and He, Conghui and Yan, Rui and Wu, Lijun},
+  year={2026}
+}
+```
 
-### Framework versions
+## License
 
-- Transformers 4.51.3
-- Pytorch 2.6.0+cu124
-- Datasets 3.5.0
-- Tokenizers 0.21.1
+This model is released under the Apache 2.0 license. The base model (Qwen3-4B-Base) is subject to its own license terms.