|
import torch |
|
from transformers import AutoModelForCausalLM, AutoTokenizer, pipeline |
|
from textwrap import dedent |
|
from huggingface_hub import login |
|
import os |
|
from dotenv import load_dotenv |
|
|
|
load_dotenv() |
|
login( |
|
token=os.environ["HF_TOKEN"], |
|
) |
|
|
|
MODEL_LIST = [ |
|
"EmergentMethods/Phi-3-mini-4k-instruct-graph", |
|
"EmergentMethods/Phi-3-mini-128k-instruct-graph", |
|
|
|
] |
|
|
|
torch.random.manual_seed(0) |
|
|
|
class Phi3InstructGraph: |
|
def __init__(self, model = "EmergentMethods/Phi-3-mini-4k-instruct-graph"): |
|
if model not in MODEL_LIST: |
|
raise ValueError(f"model must be one of {MODEL_LIST}") |
|
|
|
self.model_path = model |
|
self.model = AutoModelForCausalLM.from_pretrained( |
|
self.model_path, |
|
device_map="cuda", |
|
torch_dtype="auto", |
|
trust_remote_code=True, |
|
) |
|
self.tokenizer = AutoTokenizer.from_pretrained(self.model_path) |
|
self.pipe = pipeline( |
|
"text-generation", |
|
model=self.model, |
|
tokenizer=self.tokenizer, |
|
) |
|
|
|
def _generate(self, messages): |
|
generation_args = { |
|
"max_new_tokens": 2000, |
|
"return_full_text": False, |
|
"temperature": 0.1, |
|
"do_sample": False, |
|
} |
|
|
|
return self.pipe(messages, **generation_args) |
|
|
|
def _get_messages(self, text): |
|
context = dedent("""\n |
|
A chat between a curious user and an artificial intelligence Assistant. The Assistant is an expert at identifying entities and relationships in text. The Assistant responds in JSON output only. |
|
|
|
The User provides text in the format: |
|
|
|
-------Text begin------- |
|
<User provided text> |
|
-------Text end------- |
|
|
|
The Assistant follows the following steps before replying to the User: |
|
|
|
1. **identify the most important entities** The Assistant identifies the most important entities in the text. These entities are listed in the JSON output under the key "nodes", they follow the structure of a list of dictionaries where each dict is: |
|
|
|
"nodes":[{"id": <entity N>, "type": <type>, "detailed_type": <detailed type>}, ...] |
|
|
|
where "type": <type> is a broad categorization of the entity. "detailed type": <detailed_type> is a very descriptive categorization of the entity. |
|
|
|
2. **determine relationships** The Assistant uses the text between -------Text begin------- and -------Text end------- to determine the relationships between the entities identified in the "nodes" list defined above. These relationships are called "edges" and they follow the structure of: |
|
|
|
"edges":[{"from": <entity 1>, "to": <entity 2>, "label": <relationship>}, ...] |
|
|
|
The <entity N> must correspond to the "id" of an entity in the "nodes" list. |
|
|
|
The Assistant never repeats the same node twice. The Assistant never repeats the same edge twice. |
|
The Assistant responds to the User in JSON only, according to the following JSON schema: |
|
|
|
{"type":"object","properties":{"nodes":{"type":"array","items":{"type":"object","properties":{"id":{"type":"string"},"type":{"type":"string"},"detailed_type":{"type":"string"}},"required":["id","type","detailed_type"],"additionalProperties":false}},"edges":{"type":"array","items":{"type":"object","properties":{"from":{"type":"string"},"to":{"type":"string"},"label":{"type":"string"}},"required":["from","to","label"],"additionalProperties":false}}},"required":["nodes","edges"],"additionalProperties":false} |
|
""") |
|
|
|
user_message = dedent(f"""\n |
|
-------Text begin------- |
|
{text} |
|
-------Text end------- |
|
""") |
|
|
|
if self.model_path == "EmergentMethods/Phi-3-medium-128k-instruct-graph": |
|
|
|
messages = [ |
|
{ |
|
"role": "user", |
|
"content": f"{context}\n Input: {user_message}", |
|
} |
|
] |
|
return messages |
|
else: |
|
messages = [ |
|
{ |
|
"role": "system", |
|
"content": context |
|
}, |
|
{ |
|
"role": "user", |
|
"content": user_message |
|
} |
|
] |
|
return messages |
|
|
|
|
|
def extract(self, text): |
|
messages = self._get_messages(text) |
|
pipe_output = self._generate(messages) |
|
|
|
return pipe_output[0]["generated_text"] |
|
|