File size: 4,443 Bytes
ed5a986 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 | # PLBart: code translation with Transformers
This page is the Transformers contract of the PLBart 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 `uclanlp/plbart-java-cs`. 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.
## Translate a batch of Java methods to C#
```py runnable:translate
import torch
from transformers import AutoModelForSeq2SeqLM, AutoTokenizer, set_seed
import json # doc-builder: hide
import os # doc-builder: hide
model_id = "uclanlp/plbart-java-cs"
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) # seeds torch, NumPy, and Python: loading, dropout, and data augmentation draw random numbers
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForSeq2SeqLM.from_pretrained(model_id).to(device)
if variant: # doc-builder: hide
model = AutoModelForSeq2SeqLM.from_pretrained(model_id, **variant).to(device)
java_methods = [
"public int maximum(int a, int b, int c){return Math.Max(a, Math.Max(b, c));}",
"public int multiply(int a, int b, int c){return a * b * c;}",
"public int add(int a, int b){return a + b;}",
"public int subtract(int a, int b){return a - b;}",
]
inputs = tokenizer(java_methods, return_tensors="pt", padding=True).to(device)
generated = model.generate(**inputs, max_new_tokens=32, do_sample=False)
print(tokenizer.batch_decode(generated, skip_special_tokens=True))
```
## Decode a batch of token ids
`decode` accepts a batch and returns one string per sequence, cleaning up the spaces the tokenizer puts before punctuation.
```py runnable:translate:3
punctuated = ["public int maximum(int a, int b, int c){return Math.Max(a, Math.Max(b, c));}", "public int multiply(int a, int b, int c){return a * b * c;}"]
roundtrip = tokenizer.decode(tokenizer(punctuated).input_ids, skip_special_tokens=True)
print(roundtrip)
```
## One training step
```py runnable:translate:3
targets = [
"public int Maximum(int a, int b, int c){return Math.Max(a, Math.Max(b, c));}",
"public int Multiply(int a, int b, int c){return a * b * c;}",
"public int Add(int a, int b){return a + b;}",
"public int Subtract(int a, int b){return a - b;}",
]
labels = tokenizer(text_target=targets, return_tensors="pt", padding=True).input_ids.to(device)
labels[labels == tokenizer.pad_token_id] = -100
with torch.no_grad(): # doc-builder: hide
# Teacher-forced log-probabilities of the reference translations.
logprobs = model(**inputs, labels=labels).logits.float().log_softmax(-1)
label_logprobs = logprobs.gather(-1, labels.clamp(min=0)[..., None])[..., 0].masked_fill(labels == -100, 0.0)
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())
checked = ("model.encoder.layers.0.self_attn.q_proj.weight", "model.decoder.layers.0.encoder_attn.q_proj.weight")
result = {
"runtime": {
"device": model.device.type,
"dtype": str(model.dtype).removeprefix("torch."),
"attn_implementation": model.config._attn_implementation,
},
"generated_tokens": generated.tolist(),
"generated_text": tokenizer.batch_decode(generated, skip_special_tokens=True),
"label_logprobs": label_logprobs.tolist(),
"decoded_roundtrip": roundtrip,
"training": {
"loss": loss.item(),
"gradients": {n: bool(named[n].grad is not None and 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))
```
|