commit from nbouali

README.md

@@ -0,0 +1,36 @@
+# FlauBert finetuned on French cooking recipes
+
+This model is finetuned on a sequence classification task that associate each sequence to the appropriate recipe category. 
+
+### How to use it ?
+
+```python
+from transformers import AutoTokenizer, AutoModelForSequenceClassification
+from transformers import TextClassificationPipeline
+
+loaded_tokenizer = AutoTokenizer.from_pretrained("nbouali/flaubert-base-uncased-finetuned-cooking")
+loaded_model = AutoModelForSequenceClassification.from_pretrained("nbouali/flaubert-base-uncased-finetuned-cooking")
+
+nlp = TextClassificationPipeline(model=loaded_model,tokenizer=loaded_tokenizer,task="Recipe classification")
+
+print(nlp("Lasagnes à la bolognaise"))
+```
+
+```
+[{'label': 'LABEL_6', 'score': 0.9921900033950806}]
+```
+### Label encoding 
+
+| label | Recipe Category |
+|:------:|:--------------:|
+|     0 |'Accompagnement' |
+|     1 |  'Amuse-gueule' |
+|     2 | 'Boisson' |
+|     3 | 'Confiserie' |
+|     4 | 'Dessert'|
+|     5 | 'Entrée' |
+|     6 |'Plat principal' |
+|     7 | 'Sauce' |
+
+
+If you would like to know more about this model you can refer to [our blog post](https://medium.com/unify-data-office/a-cooking-language-model-fine-tuned-on-dozens-of-thousands-of-french-recipes-bcdb8e560571)

config.json

@@ -0,0 +1,98 @@
+{
+  "amp": 1,
+  "architectures": [
+    "FlaubertForSequenceClassification"
+  ],
+  "asm": false,
+  "attention_dropout": 0.1,
+  "bos_index": 0,
+  "bos_token_id": 0,
+  "bptt": 512,
+  "causal": false,
+  "clip_grad_norm": 5,
+  "dropout": 0.1,
+  "emb_dim": 768,
+  "embed_init_std": 0.02209708691207961,
+  "encoder_only": true,
+  "end_n_top": 5,
+  "eos_index": 1,
+  "fp16": true,
+  "gelu_activation": true,
+  "group_by_size": true,
+  "id2label": {
+    "0": "LABEL_0",
+    "1": "LABEL_1",
+    "2": "LABEL_2",
+    "3": "LABEL_3",
+    "4": "LABEL_4",
+    "5": "LABEL_5",
+    "6": "LABEL_6",
+    "7": "LABEL_7"
+  },
+  "id2lang": {
+    "0": "fr"
+  },
+  "init_std": 0.02,
+  "is_encoder": true,
+  "label2id": {
+    "LABEL_0": 0,
+    "LABEL_1": 1,
+    "LABEL_2": 2,
+    "LABEL_3": 3,
+    "LABEL_4": 4,
+    "LABEL_5": 5,
+    "LABEL_6": 6,
+    "LABEL_7": 7
+  },
+  "lang2id": {
+    "fr": 0
+  },
+  "lang_id": 0,
+  "langs": [
+    "fr"
+  ],
+  "layer_norm_eps": 1e-12,
+  "layerdrop": 0.0,
+  "lg_sampling_factor": -1,
+  "lgs": "fr",
+  "mask_index": 5,
+  "mask_token_id": 0,
+  "max_batch_size": 0,
+  "max_position_embeddings": 512,
+  "max_vocab": -1,
+  "mlm_steps": [
+    [
+      "fr",
+      null
+    ]
+  ],
+  "model_type": "flaubert",
+  "n_heads": 12,
+  "n_langs": 1,
+  "n_layers": 12,
+  "output_hidden_states": true,
+  "pad_index": 2,
+  "pad_token_id": 2,
+  "pre_norm": false,
+  "sample_alpha": 0,
+  "share_inout_emb": true,
+  "sinusoidal_embeddings": false,
+  "start_n_top": 5,
+  "summary_activation": null,
+  "summary_first_dropout": 0.1,
+  "summary_proj_to_labels": true,
+  "summary_type": "first",
+  "summary_use_proj": true,
+  "tokens_per_batch": -1,
+  "unk_index": 3,
+  "use_lang_emb": true,
+  "vocab_size": 67542,
+  "word_blank": 0,
+  "word_dropout": 0,
+  "word_keep": 0.1,
+  "word_mask": 0.8,
+  "word_mask_keep_rand": "0.8,0.1,0.1",
+  "word_pred": 0.15,
+  "word_rand": 0.1,
+  "word_shuffle": 0
+}

pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1e6bba1825513a12f9986fdc4af4db785a855d9d6f5d9b7187e7538ac73fa4f1
+size 549355707

special_tokens_map.json

@@ -0,0 +1 @@
+{"bos_token": "<s>", "unk_token": "<unk>", "sep_token": "</s>", "pad_token": "<pad>", "cls_token": "</s>", "mask_token": "<special1>", "additional_special_tokens": ["<special0>", "<special1>", "<special2>", "<special3>", "<special4>", "<special5>", "<special6>", "<special7>", "<special8>", "<special9>"]}

tokenizer_config.json

@@ -0,0 +1 @@
+{"do_lowercase": true, "do_lower_case": true, "model_max_length": 512}