{"metadata":{"kernelspec":{"language":"python","display_name":"Python 3","name":"python3"},"language_info":{"name":"python","version":"3.11.13","mimetype":"text/x-python","codemirror_mode":{"name":"ipython","version":3},"pygments_lexer":"ipython3","nbconvert_exporter":"python","file_extension":".py"},"kaggle":{"accelerator":"gpu","dataSources":[{"sourceType":"competition","sourceId":121150,"databundleVersionId":15833391},{"sourceType":"datasetVersion","sourceId":15109379,"datasetId":9674296,"databundleVersionId":15995807}],"dockerImageVersionId":31193,"isInternetEnabled":false,"language":"python","sourceType":"notebook","isGpuEnabled":true}},"nbformat_minor":4,"nbformat":4,"cells":[{"cell_type":"code","source":"import os, re\nimport pandas as pd\nimport torch\nfrom torch.utils.data import DataLoader, Dataset\nfrom transformers import AutoTokenizer, AutoModelForSeq2SeqLM\nfrom tqdm import tqdm\n\nMODEL_PATH = \"/kaggle/input/datasets/konstantinboyko/akka-2025-03-07-01-akkadian-checkpoint-5304-17\"\nTEST_DATA_PATH = \"/kaggle/input/deep-past-initiative-machine-translation/test.csv\"\nBATCH_SIZE = 8 #4\nMAX_LENGTH = 512\nDEVICE = \"cuda\" if torch.cuda.is_available() else \"cpu\"\ntokenizer = AutoTokenizer.from_pretrained(MODEL_PATH)\nmodel = AutoModelForSeq2SeqLM.from_pretrained(MODEL_PATH).to(DEVICE)\nmodel.eval()\ntest_df = pd.read_csv(TEST_DATA_PATH)\nPREFIX = \"translate Akkadian to English: \"\n\nclass InferenceDataset(Dataset):\n    def __init__(self, df, tokenizer):\n        self.texts = df['transliteration'].astype(str).tolist()\n        self.texts = [PREFIX + i for i in self.texts]\n        self.tokenizer = tokenizer\n        \n    def __len__(self):\n        return len(self.texts)\n    \n    def __getitem__(self, idx):\n        text = self.texts[idx]\n        inputs = self.tokenizer(\n            text, \n            max_length=MAX_LENGTH, \n            padding=\"max_length\", \n            truncation=True, \n            return_tensors=\"pt\"\n        )\n        return {\n            \"input_ids\": inputs[\"input_ids\"].squeeze(0),\n            \"attention_mask\": inputs[\"attention_mask\"].squeeze(0)\n        }\n\ntest_dataset = InferenceDataset(test_df, tokenizer)\ntest_loader = DataLoader(test_dataset, batch_size=BATCH_SIZE, shuffle=False)\n","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"code","source":"print(\"Starting Inference...\")\nDEVICE = torch.device(\"cuda\" if torch.cuda.is_available() else \"cpu\")\nmodel = model.to(DEVICE)\nmodel.eval()\nmodel.float()  # <-- force les poids en FP32\n# Optionnel: si tu veux être sûr de ne pas avoir de comportements bizarres \"mixed\"\n# (sur P100 TF32 n'existe pas, mais ça ne coûte rien)\nall_predictions = []\ntorch.set_grad_enabled(False)\nwith torch.inference_mode():  # mieux que no_grad pour l'inférence\n    for batch in tqdm(test_loader):\n        input_ids = batch[\"input_ids\"].to(DEVICE)              # int64, normal\n        attention_mask = batch[\"attention_mask\"].to(DEVICE)    # int64/bool, normal\n        # IMPORTANT: ne pas mettre autocast ici (pas de with torch.cuda.amp.autocast())\n        outputs = model.generate(\n            input_ids=input_ids,\n            attention_mask=attention_mask,\n            max_length=MAX_LENGTH,\n            num_beams=4,\n            early_stopping=True,\n        )\n        decoded = tokenizer.batch_decode(outputs, skip_special_tokens=True)\n        all_predictions.extend([d.strip() for d in decoded])\n        # all_predictions contient une prédiction par élément du test_loader\n        predictions = []\n        for txt in all_predictions:\n            predictions.append(txt)\n\nprint(next(model.parameters()).dtype)  # doit afficher: torch.float32\nsubmission = pd.DataFrame({\n    \"id\": test_df[\"id\"],\n    \"translation\": predictions\n})\nsubmission[\"translation\"] = submission[\"translation\"].apply(lambda x: x if len(x) > 0 else \"broken text\")\nsubmission.to_csv(\"submission.csv\", index=False)\nprint(\"Submission file saved successfully!\")\nsubmission.head()\n","metadata":{"trusted":true},"outputs":[],"execution_count":null}]}