File size: 12,462 Bytes
469ef7f
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
"""Fine-tune xlm-roberta-base for 6-class intent classification.

Intents: booking, complaint, farewell, greeting, inquiry, other.
Languages covered: AR, EN, FR (data was stratified across all of them).

Inputs:
  data/processed/intent/             (HuggingFace DatasetDict)
  data/processed/intent/labels.json

Outputs:
  models/intent_classifier/                  (best model + tokenizer)
  models/intent_classifier/eval_results.json (test metrics + classification report)
  models/intent_classifier/runs/             (training checkpoints + logs)

GPU notes for GTX 1650 (3.6 GB VRAM): same as train_lang_detector.py — we use
fp16 + gradient_checkpointing + batch=8 with grad_accum=2.

Usage:
  python src/train_intent.py
  python src/train_intent.py --epochs 3
  python src/train_intent.py --quick
"""

from __future__ import annotations

import os
# Reduce CUDA memory fragmentation on tight-VRAM GPUs (must precede torch import).
os.environ.setdefault("PYTORCH_CUDA_ALLOC_CONF", "expandable_segments:True")

import argparse
import inspect
import json
import shutil
import sys
from pathlib import Path
from typing import Any

import numpy as np
import torch
from datasets import load_from_disk
from sklearn.metrics import (
    accuracy_score, classification_report, f1_score,
    precision_score, recall_score,
)
from transformers import (
    AutoModelForSequenceClassification,
    AutoTokenizer,
    DataCollatorWithPadding,
    Trainer,
    TrainingArguments,
    set_seed,
)

PROJECT_ROOT = Path(__file__).resolve().parent.parent
DATA_DIR = PROJECT_ROOT / "data" / "processed" / "intent"
LABELS_FILE = DATA_DIR / "labels.json"
OUT_DIR = PROJECT_ROOT / "models" / "intent_classifier"
RUNS_DIR = OUT_DIR / "runs"

# Spec called for xlm-roberta-base, but at 270M params it does not fit in
# the GTX 1650's 3.6 GB VRAM during optimizer step — Adam needs ~2.2 GB of
# state and even Adafactor allocates a ~770 MB temp tensor for grad**2 over
# the embedding matrix. distilbert-base-multilingual-cased is the standard
# substitute: 134M params, trained on 104 languages (AR/EN/FR included),
# typically within 1-3% F1 of XLM-R on classification.
MODEL_NAME = "distilbert-base-multilingual-cased"
MAX_LENGTH = 128
SEED = 42


def _trainer_with_tokenizer(tokenizer, **kwargs: Any) -> Trainer:
    """Construct Trainer with whichever tokenizer kwarg this transformers
    version expects (`processing_class` in 5.x, `tokenizer` in 4.x)."""
    params = inspect.signature(Trainer.__init__).parameters
    if "processing_class" in params:
        kwargs["processing_class"] = tokenizer
    elif "tokenizer" in params:
        kwargs["tokenizer"] = tokenizer
    return Trainer(**kwargs)


def main() -> int:
    """Train XLM-R for intent classification. Returns exit code."""
    parser = argparse.ArgumentParser(description=__doc__.split("\n")[0])
    parser.add_argument("--epochs", type=int, default=5,
                        help="Number of training epochs (default 5).")
    parser.add_argument("--batch-size", type=int, default=8,
                        help="Per-device train batch size (default 8 — fits comfortably "
                             "with distilbert-multilingual on a 3.6 GB GPU).")
    parser.add_argument("--quick", action="store_true",
                        help="Sanity smoke test: 1 epoch, 500 train rows.")
    parser.add_argument("--lr", type=float, default=2e-5)
    parser.add_argument("--optim", type=str, default="adamw_torch",
                        help="Optimizer name. AdamW fits with the smaller distilbert model.")
    args = parser.parse_args()

    set_seed(SEED)
    print("=" * 72)
    print("Train intent classifier  (xlm-roberta-base, 6 classes)")
    print("=" * 72)
    print(f"  Data dir : {DATA_DIR}")
    print(f"  Out dir  : {OUT_DIR}")
    print(f"  Epochs   : {args.epochs}{' (QUICK)' if args.quick else ''}")
    print(f"  Batch    : {args.batch_size}  (effective ≈ {args.batch_size * 2} via accum)")
    print(f"  Optimizer: {args.optim}")

    # --- Labels --------------------------------------------------------------
    labels_payload = json.loads(LABELS_FILE.read_text())
    label_to_id: dict[str, int] = labels_payload["label_to_id"]
    id_to_label: dict[int, str] = {int(k): v for k, v in labels_payload["id_to_label"].items()}
    label_names = [id_to_label[i] for i in range(len(id_to_label))]
    num_labels = len(label_names)
    print(f"  Labels   : {label_to_id}")

    # --- Datasets ------------------------------------------------------------
    ds = load_from_disk(str(DATA_DIR))
    print(f"  Splits   : train={len(ds['train'])} val={len(ds['validation'])} "
          f"test={len(ds['test'])}")

    if args.quick:
        ds["train"] = ds["train"].shuffle(seed=SEED).select(range(min(500, len(ds["train"]))))
        ds["validation"] = ds["validation"].select(range(min(120, len(ds["validation"]))))
        ds["test"] = ds["test"].select(range(min(120, len(ds["test"]))))
        print(f"  QUICK    : sliced to {len(ds['train'])}/{len(ds['validation'])}/{len(ds['test'])}")

    # --- Tokenize ------------------------------------------------------------
    print("\nLoading tokenizer & model ...")
    tokenizer = AutoTokenizer.from_pretrained(MODEL_NAME)

    def tokenize(batch: dict[str, list]) -> dict[str, Any]:
        return tokenizer(batch["text"], truncation=True, max_length=MAX_LENGTH)

    drop_cols = [c for c in ds["train"].column_names if c not in ("label",)]
    ds_tok = ds.map(tokenize, batched=True, remove_columns=drop_cols,
                    desc="Tokenizing")

    # --- Model ---------------------------------------------------------------
    model = AutoModelForSequenceClassification.from_pretrained(
        MODEL_NAME,
        num_labels=num_labels,
        id2label=id_to_label,
        label2id=label_to_id,
    )

    # Free any lingering CUDA blocks before optimizer states are allocated.
    if torch.cuda.is_available():
        torch.cuda.empty_cache()

    # --- Training arguments --------------------------------------------------
    n_epochs = 1 if args.quick else args.epochs
    RUNS_DIR.mkdir(parents=True, exist_ok=True)

    training_args_kwargs = dict(
        output_dir=str(RUNS_DIR),
        num_train_epochs=n_epochs,
        per_device_train_batch_size=args.batch_size,
        per_device_eval_batch_size=args.batch_size * 2,  # eval has no grads -> larger ok
        gradient_accumulation_steps=2,                    # effective batch = batch * 2 = 16 (matches spec)
        optim=args.optim,
        learning_rate=args.lr,
        warmup_steps=100 if not args.quick else 10,
        weight_decay=0.01,
        fp16=True,
        gradient_checkpointing=True,
        eval_strategy="epoch",
        save_strategy="epoch",
        save_total_limit=1,            # keep only the current-best checkpoint (was 2)
        save_only_model=True,          # skip optimizer/scheduler state — saves ~340 MB/ckpt
        load_best_model_at_end=True,
        metric_for_best_model="f1_macro",  # macro is more sensitive to minority classes
        greater_is_better=True,
        logging_steps=50,
        report_to="none",
        dataloader_num_workers=0,
        seed=SEED,
    )
    try:
        training_args = TrainingArguments(**training_args_kwargs)
    except TypeError:
        training_args_kwargs["evaluation_strategy"] = training_args_kwargs.pop("eval_strategy")
        training_args = TrainingArguments(**training_args_kwargs)

    # --- Metrics -------------------------------------------------------------
    def compute_metrics(eval_pred) -> dict[str, float]:
        logits, labels = eval_pred
        if isinstance(logits, tuple):
            logits = logits[0]
        preds = np.argmax(logits, axis=-1)
        return {
            "accuracy": accuracy_score(labels, preds),
            "f1": f1_score(labels, preds, average="weighted", zero_division=0),
            "f1_macro": f1_score(labels, preds, average="macro", zero_division=0),
            "precision": precision_score(labels, preds, average="weighted", zero_division=0),
            "recall": recall_score(labels, preds, average="weighted", zero_division=0),
        }

    # --- Trainer -------------------------------------------------------------
    trainer = _trainer_with_tokenizer(
        tokenizer,
        model=model,
        args=training_args,
        train_dataset=ds_tok["train"],
        eval_dataset=ds_tok["validation"],
        data_collator=DataCollatorWithPadding(tokenizer),
        compute_metrics=compute_metrics,
    )

    # --- Train ---------------------------------------------------------------
    print("\nStarting training ...")
    train_result = trainer.train()
    print(f"  ✓ training done. final loss = {train_result.training_loss:.4f}")

    # --- Save best model -----------------------------------------------------
    OUT_DIR.mkdir(parents=True, exist_ok=True)
    trainer.save_model(str(OUT_DIR))
    tokenizer.save_pretrained(str(OUT_DIR))
    shutil.copy(LABELS_FILE, OUT_DIR / "labels.json")

    # --- Final evaluation on test set ---------------------------------------
    print("\nEvaluating on TEST split ...")
    test_metrics = trainer.evaluate(ds_tok["test"], metric_key_prefix="test")
    test_pred = trainer.predict(ds_tok["test"])
    if isinstance(test_pred.predictions, tuple):
        test_logits = test_pred.predictions[0]
    else:
        test_logits = test_pred.predictions
    pred_ids = np.argmax(test_logits, axis=-1)
    true_ids = test_pred.label_ids

    report_dict = classification_report(
        true_ids, pred_ids,
        labels=list(range(num_labels)),
        target_names=label_names,
        output_dict=True, zero_division=0,
    )
    report_text = classification_report(
        true_ids, pred_ids,
        labels=list(range(num_labels)),
        target_names=label_names,
        zero_division=0,
    )
    print("\nClassification report on TEST:")
    print(report_text)

    # --- Per-language breakdown ---------------------------------------------
    # Reload the test split with language column to break down errors by language
    test_with_lang = load_from_disk(str(DATA_DIR))["test"]
    if args.quick:
        test_with_lang = test_with_lang.select(range(min(120, len(test_with_lang))))
    per_lang: dict[str, dict[str, float]] = {}
    if "language" in test_with_lang.column_names:
        languages = test_with_lang["language"]
        for lang in sorted(set(languages)):
            mask = np.array([la == lang for la in languages])
            if not mask.any():
                continue
            lp = pred_ids[mask]
            lt = true_ids[mask]
            per_lang[lang] = {
                "n": int(mask.sum()),
                "accuracy": float(accuracy_score(lt, lp)),
                "f1_weighted": float(f1_score(lt, lp, average="weighted", zero_division=0)),
                "f1_macro": float(f1_score(lt, lp, average="macro", zero_division=0)),
            }
        print("\nPer-language metrics on TEST:")
        for lang, m in per_lang.items():
            print(f"  {lang}: n={m['n']}  acc={m['accuracy']:.4f}  "
                  f"f1_w={m['f1_weighted']:.4f}  f1_m={m['f1_macro']:.4f}")

    # --- Save eval_results.json ---------------------------------------------
    payload = {
        "model_name": MODEL_NAME,
        "task": "intent",
        "num_labels": num_labels,
        "labels": label_to_id,
        "test_metrics": {k: float(v) for k, v in test_metrics.items()
                         if isinstance(v, (int, float))},
        "classification_report": report_dict,
        "per_language": per_lang,
        "training": {
            "epochs": n_epochs,
            "per_device_batch": args.batch_size,
            "grad_accum": 2,
            "effective_batch": args.batch_size * 2,
            "learning_rate": args.lr,
            "warmup_steps": training_args_kwargs.get("warmup_steps"),
            "fp16": True,
            "final_train_loss": float(train_result.training_loss),
        },
    }
    (OUT_DIR / "eval_results.json").write_text(
        json.dumps(payload, indent=2, ensure_ascii=False)
    )
    print(f"\n✓ Saved model to {OUT_DIR}")
    print(f"✓ Saved eval_results.json to {OUT_DIR / 'eval_results.json'}")
    return 0


if __name__ == "__main__":
    try:
        sys.exit(main())
    except KeyboardInterrupt:
        print("\nAborted by user.")
        sys.exit(130)