| # ModernBERT: fill-mask with Transformers |
|
|
| This page is the Transformers contract of the ModernBERT MIC and also a usage example. [doc-builder](https://github.com/huggingface/doc-builder/blob/main/docs/runnable-code-blocks.md) renders the `runnable` blocks below as documentation, and its pytest plugin runs them as written against `answerdotai/ModernBERT-base`. The MIC runner runs the same blocks against each slim fixture and each checkpoint in [`integration.json`](../../integration.json). Lines marked `# doc-builder: hide` only matter there: they swap in the pinned checkpoint or variant and print the result record. |
|
|
| ## Predict the masked word |
|
|
| ```py runnable:fill_mask |
| import torch |
| from transformers import AutoModelForMaskedLM, AutoTokenizer, set_seed |
| import json # doc-builder: hide |
| import os # doc-builder: hide |
| |
| model_id = "answerdotai/ModernBERT-base" |
| model_id = os.environ.get("MIC_MODEL", model_id) # doc-builder: hide |
| variant = json.loads(os.environ.get("MIC_MODEL_KWARGS", "{}")) # doc-builder: hide |
| device = "cuda" if torch.cuda.is_available() else "cpu" |
| set_seed(0) |
| |
| tokenizer = AutoTokenizer.from_pretrained(model_id) |
| model = AutoModelForMaskedLM.from_pretrained(model_id).to(device) |
| if variant: # doc-builder: hide |
| model = AutoModelForMaskedLM.from_pretrained(model_id, **variant).to(device) |
| |
| prompts = [ |
| "The capital of France is [MASK].", |
| "Water is made of hydrogen and [MASK].", |
| ] |
| inputs = tokenizer(prompts, return_tensors="pt", padding=True).to(device) |
| |
| model.eval() |
| with torch.no_grad(): |
| logits = model(**inputs).logits |
| masked = inputs["input_ids"] == tokenizer.mask_token_id |
| predicted = logits[masked].argmax(dim=-1) |
| print(tokenizer.convert_ids_to_tokens(predicted)) |
| ``` |
|
|
| ## One training step |
|
|
| ```py runnable:fill_mask:2 |
| answers = ["paris", "oxygen"] |
| answer_ids = tokenizer.convert_tokens_to_ids(answers) |
| labels = torch.full_like(inputs["input_ids"], -100) |
| labels[masked] = torch.tensor(answer_ids, device=device) |
| |
| model.train() |
| optimizer = torch.optim.SGD(model.parameters(), lr=0.01) |
| loss = model(**inputs, labels=labels).loss |
| loss.backward() |
| before = {name: p.detach().clone() for name, p in model.named_parameters()} # doc-builder: hide |
| optimizer.step() |
| print(f"loss: {loss.item():.4f}") |
| |
| if os.environ.get("MIC_RUNNER"): # doc-builder: hide |
| named = dict(model.named_parameters()) |
| grad_params = [n for n, p in named.items() if p.grad is not None] |
| checked = grad_params[:2] |
| result = { |
| "runtime": { |
| "device": model.device.type, |
| "dtype": str(model.dtype).removeprefix("torch."), |
| "attn_implementation": model.config._attn_implementation, |
| }, |
| "predicted_tokens": predicted.tolist(), |
| "predicted_text": tokenizer.convert_ids_to_tokens(predicted), |
| "answer_logits": logits[masked][:, answer_ids].float().tolist(), |
| "training": { |
| "loss": loss.item(), |
| "gradients": {n: bool(torch.count_nonzero(named[n].grad) > 0 and torch.isfinite(named[n].grad).all()) for n in checked}, |
| "updated": any(not torch.equal(p, before[n]) for n, p in named.items()), |
| }, |
| } |
| print(json.dumps(result)) |
| ``` |
|
|
