First commit

README.md

@@ -0,0 +1,56 @@
+---
+
+pipeline_tag: text-classification
+inference: false
+language: pt
+tags:
+- transformers
+
+---
+
+# Prompsit/paraphrase-bert-pt
+
+This model allows to evaluate paraphrases for a given phrase.  
+
+We have fine-tuned this model from pretrained "neuralmind/bert-base-portuguese-cased".
+
+Model built under a TSI-100905-2019-4 project, co-financed by Ministry of Economic Affairs and Digital Transformation from the Government of Spain.
+
+# How to use it
+
+The model answer the following question: Is "phrase B" a paraphrase of "phrase A".
+
+Please note that we're considering phrases instead of sentences. Therefore, we must take into account that the model doesn't expect to find punctuation marks or long pieces of text.
+
+Resulting probabilities correspond to classes:  
+
+* 0: Not a paraphrase
+* 1: It's a paraphrase
+
+So, considering the phrase "logo após o homicídio" and a candidate paraphrase like "pouco depois do assassinato", you can use the model like this:
+
+```
+
+import torch
+
+from transformers import AutoTokenizer, AutoModelForSequenceClassification
+tokenizer = AutoTokenizer.from_pretrained("Prompsit/paraphrase-bert-pt")
+model = AutoModelForSequenceClassification.from_pretrained("Prompsit/paraphrase-bert-pt")
+
+input = tokenizer('logo após o homicídio','pouco depois do assassinato',return_tensors='pt')
+logits = model(**input).logits
+soft = torch.nn.Softmax(dim=1)
+print(soft(logits))
+
+```
+
+Code output is:
+
+ ``` 
+
+ tensor([[0.2137, 0.7863]], grad_fn=<SoftmaxBackward>)
+
+ ```
+
+As the probability of 1 (=It's a paraphrase) is 0.7863 and the probability of 0 (=It is not a paraphrase) is 0.2137, we can conclude, for our previous example, that "pouco depois do assassinato" is a paraphrase of "logo após o homicidio".
+

config.json

@@ -0,0 +1,41 @@
+{
+  "_name_or_path": "neuralmind/bert-base-portuguese-cased",
+  "architectures": [
+    "BertForSequenceClassification"
+  ],
+  "attention_probs_dropout_prob": 0.1,
+  "classifier_dropout": null,
+  "directionality": "bidi",
+  "hidden_act": "gelu",
+  "hidden_dropout_prob": 0.1,
+  "hidden_size": 768,
+  "id2label": {
+    "0": "Not Paraphrase",
+    "1": "Paraphrase"
+  },
+  "initializer_range": 0.02,
+  "intermediate_size": 3072,
+  "label2id": {
+    "Not Paraphrase": 0,
+    "Paraphrase": 1
+  },
+  "layer_norm_eps": 1e-12,
+  "max_position_embeddings": 512,
+  "model_type": "bert",
+  "num_attention_heads": 12,
+  "num_hidden_layers": 12,
+  "output_past": true,
+  "pad_token_id": 0,
+  "pooler_fc_size": 768,
+  "pooler_num_attention_heads": 12,
+  "pooler_num_fc_layers": 3,
+  "pooler_size_per_head": 128,
+  "pooler_type": "first_token_transform",
+  "position_embedding_type": "absolute",
+  "problem_type": "single_label_classification",
+  "torch_dtype": "float32",
+  "transformers_version": "4.11.3",
+  "type_vocab_size": 2,
+  "use_cache": true,
+  "vocab_size": 29794
+}

pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bff4e0b43c243a7e7668cee0be15c0e062f237164ed8393bfac8d5ad73397f15
+size 435782829

special_tokens_map.json

@@ -0,0 +1 @@
+{"unk_token": "[UNK]", "sep_token": "[SEP]", "pad_token": "[PAD]", "cls_token": "[CLS]", "mask_token": "[MASK]"}

tokenizer_config.json

@@ -0,0 +1 @@
+{"do_lower_case": false, "unk_token": "[UNK]", "sep_token": "[SEP]", "pad_token": "[PAD]", "cls_token": "[CLS]", "mask_token": "[MASK]", "tokenize_chinese_chars": true, "strip_accents": null, "special_tokens_map_file": "/home/mplaza/.cache/huggingface/transformers/eecc45187d085a1169eed91017d358cc0e9cbdd5dc236bcd710059dbf0a2f816.dd8bd9bfd3664b530ea4e645105f557769387b3da9f79bdb55ed556bdd80611d", "name_or_path": "neuralmind/bert-base-portuguese-cased", "do_basic_tokenize": true, "never_split": null, "tokenizer_class": "BertTokenizer"}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0a5bc90de40c0c9489ac705d5176d0991e203c34e0c667b2630a0bdcdffe6854
+size 2799