app + chackpoint

app.py

@@ -0,0 +1,49 @@
+import gradio as gr
+from transformers import (T5Tokenizer, 
+                          T5ForConditionalGeneration, 
+                          AddedToken,
+)
+
+
+tokenizer = T5Tokenizer.from_pretrained(f"google/flan-t5-base")
+tokenizer.add_special_tokens({"additional_special_tokens": [AddedToken("\n")]})
+
+# load model
+model_cktp = "model_checkpoint"
+model = T5ForConditionalGeneration.from_pretrained(model_cktp)
+
+def predict(input):
+    input_ids = tokenizer.encode(input, return_tensors="pt")
+    outputs = model.generate(input_ids, 
+                            max_length=200,
+                            early_stopping=True)
+    response = tokenizer.decode(outputs[0], skip_special_tokens=False)
+    response = response.replace("<pad>", "").replace("</s>", "")
+    return response
+
+
+examples = [["""Service: telecom customer service.
+Customer utterance : "I'm trying to find out when my tv service will be turn back on??????"|
+Extract all strictly unnecessary sequences for the service provider to process the request/issue and then classify them using relational tags."""],
+["""Service: airline customer service.
+Customer utterance : "I need a ticket to Boston this Saturday, my son is graduating!"|
+Extract all strictly unnecessary sequences for the service provider to process the request/issue and classify them using relational tags."""]
+            ]
+
+description = """This model detects and classifies relational strategies in customer service requests, using an instruction-based approach."""
+
+demo = gr.Interface(fn=predict, 
+                    inputs="text", 
+                    outputs="text",
+                    title="FLAN-T5: Detect and Classify Relational Strategies",
+                    examples=examples,
+                    description=description)
+
+if __name__ == "__main__":
+    demo.launch() 
+    
+
+
+
+
+

model_checkpoint/config.json

@@ -0,0 +1,61 @@
+{
+  "_name_or_path": "google/flan-t5-base",
+  "architectures": [
+    "T5ForConditionalGeneration"
+  ],
+  "d_ff": 2048,
+  "d_kv": 64,
+  "d_model": 768,
+  "decoder_start_token_id": 0,
+  "dense_act_fn": "gelu_new",
+  "dropout_rate": 0.1,
+  "eos_token_id": 1,
+  "feed_forward_proj": "gated-gelu",
+  "initializer_factor": 1.0,
+  "is_encoder_decoder": true,
+  "is_gated_act": true,
+  "layer_norm_epsilon": 1e-06,
+  "model_type": "t5",
+  "n_positions": 512,
+  "num_decoder_layers": 12,
+  "num_heads": 12,
+  "num_layers": 12,
+  "output_past": true,
+  "pad_token_id": 0,
+  "relative_attention_max_distance": 128,
+  "relative_attention_num_buckets": 32,
+  "task_specific_params": {
+    "summarization": {
+      "early_stopping": true,
+      "length_penalty": 2.0,
+      "max_length": 200,
+      "min_length": 30,
+      "no_repeat_ngram_size": 3,
+      "num_beams": 4,
+      "prefix": "summarize: "
+    },
+    "translation_en_to_de": {
+      "early_stopping": true,
+      "max_length": 300,
+      "num_beams": 4,
+      "prefix": "translate English to German: "
+    },
+    "translation_en_to_fr": {
+      "early_stopping": true,
+      "max_length": 300,
+      "num_beams": 4,
+      "prefix": "translate English to French: "
+    },
+    "translation_en_to_ro": {
+      "early_stopping": true,
+      "max_length": 300,
+      "num_beams": 4,
+      "prefix": "translate English to Romanian: "
+    }
+  },
+  "tie_word_embeddings": false,
+  "torch_dtype": "float32",
+  "transformers_version": "4.26.0",
+  "use_cache": true,
+  "vocab_size": 32101
+}

model_checkpoint/generation_config.json

@@ -0,0 +1,7 @@
+{
+  "_from_model_config": true,
+  "decoder_start_token_id": 0,
+  "eos_token_id": 1,
+  "pad_token_id": 0,
+  "transformers_version": "4.26.0"
+}

model_checkpoint/pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:21fa8993f500e9404f1df4f4c116bae425d52e1be037f62bd474a94227e3bdbb
+size 990242997

model_checkpoint/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b8406259d6fb5df1141b5e09507c2ff97c99a753f3a9eac4ddbba8a3330124a4
+size 3643