Spaces:
Sleeping
Sleeping
| import gradio as gr | |
| import json | |
| from FlagEmbedding import BGEM3FlagModel | |
| model = BGEM3FlagModel('BAAI/bge-m3', use_fp16=True) | |
| with open("charters.json", "r") as f: | |
| charters = json.loads(f.read()) | |
| def greet(question): | |
| sentences = cleanSplit(charters['Boston']['current']) | |
| question_embeddings = model.encode(question, | |
| batch_size=12, | |
| max_length=128, # If you don't need such a long length, you can set a smaller value to speed up the encoding process. | |
| return_dense=False, return_sparse=False, return_colbert_vecs=True) | |
| embeddings_2 = model.encode(sentences, | |
| batch_size=12, | |
| max_length=512, | |
| return_dense=False, return_sparse=False, return_colbert_vecs=True) | |
| print(embeddings_2) | |
| return "yes" | |
| def cleanSplit(text): | |
| resp = [] | |
| docs = text.replace(".\n", ". ").split(". ") | |
| for doc in docs: | |
| if bool(doc) and len([x for x in doc if x.isalnum()]) > 0: | |
| if len(resp) > 0 and (len(resp[-1]) + len(doc)) <= 128: | |
| resp[-1] += doc.replace("\n", " ") + '. ' | |
| else: | |
| doc += '. ' | |
| resp.append(doc) | |
| return resp | |
| iface = gr.Interface(fn=greet, inputs="text", outputs="text") | |
| iface.launch() |