Initial model

README.md

@@ -0,0 +1,13 @@
+# Sentiment Analysis in English
+## bertweet-sentiment-analysis
+
+Repository: [https://github.com/finiteautomata/pysentimiento/](https://github.com/finiteautomata/pysentimiento/)
+
+
+Model trained with SemEval 2017 corpus (around ~40k tweets). Base model is [BerTweet](https://github.com/VinAIResearch/BERTweet), a RoBERTa model trained on English tweets.
+
+Uses `POS`, `NEG`, `NEU` labels.
+
+**Coming soon**: a brief paper describing the model and training.
+
+Enjoy! 🤗

added_tokens.json

@@ -0,0 +1 @@
+{"<mask>": 64000}

config.json

@@ -0,0 +1,38 @@
+{
+  "_name_or_path": "vinai/bertweet-base",
+  "architectures": [
+    "RobertaForSequenceClassification"
+  ],
+  "attention_probs_dropout_prob": 0.1,
+  "bos_token_id": 0,
+  "eos_token_id": 2,
+  "gradient_checkpointing": false,
+  "hidden_act": "gelu",
+  "hidden_dropout_prob": 0.1,
+  "hidden_size": 768,
+  "id2label": {
+    "0": "NEG",
+    "1": "NEU",
+    "2": "POS"
+  },
+  "initializer_range": 0.02,
+  "intermediate_size": 3072,
+  "label2id": {
+    "NEG": 0,
+    "NEU": 1,
+    "POS": 2
+  },
+  "layer_norm_eps": 1e-05,
+  "max_position_embeddings": 130,
+  "model_type": "roberta",
+  "num_attention_heads": 12,
+  "num_hidden_layers": 12,
+  "pad_token_id": 1,
+  "position_embedding_type": "absolute",
+  "problem_type": "single_label_classification",
+  "tokenizer_class": "BertweetTokenizer",
+  "transformers_version": "4.6.1",
+  "type_vocab_size": 1,
+  "use_cache": true,
+  "vocab_size": 64001
+}

pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ba02cf51728b37105058e26f4571f28957fd5f4116bcf00d19e06107042e3a66
+size 539694523

special_tokens_map.json

@@ -0,0 +1 @@
+{"bos_token": "<s>", "eos_token": "</s>", "unk_token": "<unk>", "sep_token": "</s>", "pad_token": "<pad>", "cls_token": "<s>", "mask_token": "<mask>"}

tokenizer_config.json

@@ -0,0 +1 @@
+{"normalization": false, "bos_token": "<s>", "eos_token": "</s>", "sep_token": "</s>", "cls_token": "<s>", "unk_token": "<unk>", "pad_token": "<pad>", "mask_token": "<mask>", "model_max_length": 128, "special_tokens_map_file": null, "tokenizer_file": null, "name_or_path": "vinai/bertweet-base"}