Ml Training RecipesSAFE

🦞+🔬 NanoResearch: The Autonomous AI Research Assistant

The instruction file, verbatim from the audited commit — this is the text the model reads, and the surface the audit's instruction layer examines. Quoted here so you can judge it without cloning anything.

---
name: ml-training-recipes
description: Battle-tested PyTorch training recipes for all domains — LLMs, vision, diffusion, medical imaging, protein/drug discovery, spatial omics, genomics. Covers training loops, optimizer selection (AdamW, Muon), LR scheduling, mixed precision, debugging, and systematic experimentation. Use when training or fine-tuning neural networks, debugging loss spikes or OOM, choosing architectures, or optimizing GPU throughput.
version: 1.0.0
author: dailycafi
license: MIT
tags: [PyTorch, Training, Optimization, LLM, Vision, Diffusion, Biomedical, Muon, AdamW, Debugging]
dependencies: [torch>=2.0.0]
---

# ML Training Recipes

Battle-tested patterns for PyTorch training across domains. Drawn from production codebases
(Karpathy's autoresearch/nanochat, torchvision, HuggingFace) and modern training practice.

## Reference files (read when needed)

- `references/architecture.md` — Transformer/LLM architecture code patterns, weight init
- `references/optimizers.md` — Muon, AdamW hybrid, per-group LR, compiled optimizer steps
- `references/domain-specific.md` — Vision, diffusion, contrastive, distributed, checkpointing, data loading
- `references/scaling-and-selection.md` — Scaling laws, compute budget tables, decision trees, DGX Spark
- `references/biomedical.md` — Drug discovery, protein models, medical imaging, genomics, clinical NLP
- `references/experiment-loop.md` — Autonomous experiment loop (autoresearch keep/discard/revert)

---

## Architecture Selection

Pick the right model by **data type** and **data scale**:

| Data Type | < 10K samples | 10K-100K | > 100K |
|-----------|--------------|----------|--------|
| **Images** | Pretrained CNN + fine-tune | Fine-tune ViT or CNN | ViT from scratch |
| **Text (gen)** | Few-shot prompting | Fine-tune GPT/LLaMA (LoRA) | Pretrain from scratch |
| **Tabular** | XGBoost/LightGBM | Still XGBoost | Neural viable |
| **Audio** | Pretrained Whisper | Fine-tune AST | Train from scratch |
| **Molecules** | Pretrained GNN | Fine-tune molecular LM | Train GNN from scratch |
| **Proteins** | ESM-2 embeddings + head | Fine-tune ESM-2 | Train protein LM |
| **Medical img** | Pretrained CNN | nnU-Net (auto-config) | Swin-UNETR / MedSAM |

**Key principle**: architecture matters less than training recipe at equal compute. A well-tuned
ResNet beats a poorly-tuned ViT (ref: "ResNet Strikes Back", Wightman 2021).

For biomedical domains, see `references/biomedical.md`.
For sequence model selection and compute planning, see `references/scaling-and-selection.md`.

---

## Scaling Laws

### Chinchilla rule (Hoffmann et al., 2022)

Compute-optimal training: **~20 tokens per parameter**.

| Model Size | Compute-Optimal | Inference-Optimal (100×) |
|-----------|----------------|--------------------------|
| 125M | 2.5B tokens | 12.5B tokens |
| 1B | 20B tokens | 100B tokens |
| 7B | 140B tokens | 700B tokens |

**FLOPs ≈ 6 × N × D** (N=params, D=tokens). Data repetition limit: ~4 epochs before diminishing returns.

---

## Training Loop

```python
import gc, time, torch

torch.manual_seed(42)
torch.set_float32_matmul_precision("high")  # TF32 on Ampere+
autocast_ctx = torch.amp.autocast(device_type="cuda", dtype=torch.bfloat16)

grad_accum_steps = total_batch_size // (batch_size * seq_len)
step = 0

while not done:
    t0 = time.time()
    for micro_step in range(grad_accum_steps):
        with autocast_ctx:
            loss = model(x, y)
        (loss / grad_accum_steps).backward()
        x, y = next(train_loader)

    update_lr(optimizer, progress)
    optimizer.step()
    model.zero_grad(set_to_none=True)  # frees memory vs zeroing

    if loss.item() > 100:  # fast-fail on divergence
        print("FAIL: loss exploded"); exit(1)

    torch.cuda.synchronize()
    if step == 0:
        gc.collect(); gc.freeze(); gc.disable()  # avoid ~500ms GC stalls
    step += 1
```

### Key principles

- **Gradient clipping**: `clip_grad_norm_(params, 1.0)` — near-universal for Transformers.
  Exception: Muon optimizer normalizes updates via orthogonalization, so clipping is optional.
- **Tensor Core alignment**: batch size, hidden dims should be multiples of 8 (bf16) or 64 (A100).
- **Time-based budgets** make experiments comparable across hardware.
- **`cudnn.benchmark = True`** for fixed-size vision inputs.

---

## Optimizer Configuration

Modern LLM training uses different optimizers per parameter group:

| Parameter Type | Optimizer | LR (base) | Weight Decay |
|---------------|-----------|-----------|--------------|
| 2D weight matrices | Muon | 0.04 | 0.2 |
| Token embeddings | AdamW | 0.6 × scale | 0.0 |
| Unembedding (lm_head) | AdamW | 0.004 × scale | 0.0 |
| Per-layer scalars | AdamW | 0.005 × scale | 0.0 |

**LR scaling by dimension**: `lr * (d_model / 768)^(-0.5)` — keeps dynamics stable across sizes.

### Rules of thumb

- Embeddings need higher LR (sparse updates). Never weight-decay embeddings.
- Weight decay scheduling: linearly decay WD to 0 over training.
- AdamW defaults: β1=0.9, β2=0.95, eps=1e-10 (not default 1e-8 — prevents stale updates in bf16).

For Muon details (polar express orthogonalization, NorMuon), see `references/optimizers.md`.

---

## Learning Rate Scheduling

### Time-based (autoresearch style)

```python
def get_lr_multiplier(progress):  # progress = elapsed_time / time_budget
    if progress < warmup_ratio:
        return progress / warmup_ratio
    elif progress < 1.0 - warmdown_ratio:
        return 1.0
    else:
        cooldown = (1.0 - progress) / warmdown_ratio
        return cooldown + (1 - cooldown) * final_lr_frac
```

### Cosine decay

```python
def get_lr(step, total_steps, max_lr, min_lr, warmup_steps):
    if step < warmup_steps:
        return max_lr * step / warmup_steps
    progress = (step - warmup_steps) / (total_steps - warmup_steps)
    return min_lr + 0.5 * (max_lr - min_lr) * (1 + math.cos(math.pi * progress))
```

**WSD (Warmup-Stable-Decay)**: gaining traction — easier to resume trainin

SAFEgrade B · trust 89/100 Nothing in the source contradicts what it says it does. Grade A is reserved for packages that have also passed the behavioural sandbox.

What the source does

Filesystem

none-observed

Network

none-observed

Shell

none-observed

Dependencies

pinned

Secrets in source

none-found

Findings (0)

No findings outside the package's declared scope.

Gates applied: no_behavioural_pass.