Initial upload: MitoSeqGen baseline checkpoint (Model 1, CE baseline)

README.md

@@ -0,0 +1,105 @@
+---
+license: unknown
+tags:
+  - biology
+  - genomics
+  - mrna
+  - codon-optimization
+  - mitochondria
+  - transformer
+  - pytorch
+library_name: pytorch
+---
+
+# MitoSeqGen — Mitochondrial Codon-Aware Sequence Generator
+
+MitoSeqGen is a sequence-to-sequence Transformer that generates mitochondrial coding
+sequences (CDS) from an input amino-acid (protein) sequence, respecting the
+vertebrate mitochondrial genetic code and codon usage patterns learned from
+mitochondrial genomes.
+
+This checkpoint is the **cross-entropy baseline** model (referred to as "Model 1" in
+the project's internal comparisons) — of the variants evaluated so far it has the
+best BLEU / mt-CAI tradeoff and the lowest GC-content drift from natural sequences.
+
+- Repo / training code: https://github.com/Maheshbonthada/MItoDNA (private)
+
+## Model details
+
+- **Architecture:** encoder-decoder Transformer (`MitoSeqTransformer`), amino-acid
+  sequence in, codon sequence out.
+- **Parameters:** ~25.2M
+- **d_model:** 384 · **heads:** 6 · **encoder/decoder layers:** 6 each ·
+  **feedforward dim:** 1536 · **dropout:** 0.1 · **max position embeddings:** 768
+- **Training objective:** token-level cross-entropy over the codon vocabulary
+- **Training data:** vertebrate mitochondrial genome CDS records, QC-filtered,
+  deduplicated, and split by phylogeny into train/val/test (see the training repo's
+  `data/` and `src/data/` for the pipeline)
+- **Epochs trained:** 30 (this checkpoint is the best-validation-loss snapshot, epoch 20)
+- **Final validation loss:** 0.901 (train loss 0.940 at epoch 30)
+- **Hardware used for training:** single NVIDIA RTX 3050 (8.6GB VRAM), bf16 mixed precision
+
+## Evaluation (n=200 held-out test proteins)
+
+| Metric | MitoSeqGen (this model) |
+|---|---|
+| mean mt-CAI | 0.870 |
+| genetic-code compliance rate | 1.00 |
+| mean BLEU vs. natural CDS | 0.313 |
+| mean GC-content delta from natural | 0.042 |
+| mean MFE delta from natural | 34.3 |
+| novel-sequence rate | 1.00 |
+
+mt-CAI = mitochondrial codon adaptation index; MFE = minimum free energy (RNA
+secondary structure, via ViennaRNA). Compared against lookup-table, random-synonymous,
+most-frequent-codon, and CodonTransformer-remap baselines in the source repo's
+evaluation reports.
+
+## Files
+
+- `pytorch_model.pt` — inference-only checkpoint: `{"model_state_dict", "config",
+  "epoch", "val_loss"}`. Optimizer/scheduler state was stripped (not needed for
+  inference); this is **not** a drop-in replacement for resuming training.
+- `config.json` — the full training config (data paths, model hyperparameters,
+  training hyperparameters, hardware settings) for this run.
+
+## Usage
+
+Requires the `MitoSeqTransformer` class and vocabularies from the training repo
+(`src/models/transformer.py`, `src/genetic_codes.py`). This checkpoint does not
+include a HF `transformers`-compatible wrapper — load it directly with PyTorch:
+
+```python
+import torch
+from src.models.transformer import MitoSeqTransformer
+from src.genetic_codes import AA_VOCAB, VOCAB  # from the training repo
+
+ckpt = torch.load("pytorch_model.pt", map_location="cpu")
+cfg = ckpt["config"]["model"]
+
+model = MitoSeqTransformer(
+    src_vocab_size=len(AA_VOCAB),
+    tgt_vocab_size=len(VOCAB),
+    d_model=cfg["d_model"],
+    nhead=cfg["nhead"],
+    num_encoder_layers=cfg["num_encoder_layers"],
+    num_decoder_layers=cfg["num_decoder_layers"],
+    dim_feedforward=cfg["dim_feedforward"],
+    dropout=cfg["dropout"],
+    max_position_embeddings=cfg["max_position_embeddings"],
+)
+model.load_state_dict(ckpt["model_state_dict"])
+model.eval()
+
+# then use src.models.generate.generate_cds(model, protein_sequence, device="cpu")
+```
+
+## Limitations
+
+- Research checkpoint, not benchmarked against a large external test set.
+- Two multi-objective variants (GC-content / mt-CAI regularized) were trained
+  alongside this baseline; one regressed on held-out evaluation and a fourth
+  (synonym-class-restricted) has not yet been fully evaluated. This baseline was
+  selected as the best-performing checkpoint among those evaluated so far, not
+  necessarily the final model for the project.
+- License unset — treat as all-rights-reserved until the repo owner adds one.

config.json

@@ -0,0 +1,41 @@
+{
+  "project": {
+    "name": "MitoSeqGen",
+    "seed": 42
+  },
+  "data": {
+    "raw_dir": "data/raw",
+    "processed_dir": "data/processed",
+    "qc_dir": "data/qc_reports",
+    "splits_dir": "data/splits",
+    "tokenized_file": "mito_cds_tokenized_augmented.json",
+    "train_split": "train_augmented.json",
+    "val_split": "val.json",
+    "test_split": "test.json"
+  },
+  "model": {
+    "d_model": 384,
+    "nhead": 6,
+    "num_encoder_layers": 6,
+    "num_decoder_layers": 6,
+    "dim_feedforward": 1536,
+    "dropout": 0.1,
+    "max_position_embeddings": 768
+  },
+  "training": {
+    "batch_size": 24,
+    "gradient_accumulation_steps": 2,
+    "learning_rate": 0.0003,
+    "weight_decay": 0.01,
+    "epochs": 30,
+    "warmup_steps": 1000,
+    "early_stopping_patience": 10,
+    "grad_clip_norm": 1.0
+  },
+  "hardware": {
+    "device": "auto",
+    "mixed_precision": "bf16",
+    "num_workers": 0,
+    "pin_memory": true
+  }
+}

pytorch_model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:895b6e24a2e0919ecd1a86bc598d1a7277e09dcfac13e26c9b438969e49fbcbf
+size 100879726