File size: 9,187 Bytes
f189f6c
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
from typing import Dict, List, Any
import os
import torch
from transformers import AutoTokenizer, AutoModel
import pandas as pd
import time
import numpy as np

class EndpointHandler:
    def __init__(self, path="insilicomedicine/precious3-gpt"):

        self.model = AutoModel.from_pretrained(path, trust_remote_code=True).to('cuda')
        self.tokenizer = AutoTokenizer.from_pretrained(path, trust_remote_code=True)
        self.model.config.pad_token_id = self.tokenizer.pad_token_id
        self.model.config.bos_token_id = self.tokenizer.bos_token_id
        self.model.config.eos_token_id = self.tokenizer.eos_token_id
        
        unique_entities_p3 = pd.read_csv('https://huggingface.co/insilicomedicine/precious3-gpt/raw/main/all_entities_with_type.csv')
        self.unique_compounds_p3 = [i.strip() for i in unique_entities_p3[unique_entities_p3.type=='compound'].entity.to_list()]
        self.unique_genes_p3 = [i.strip() for i in unique_entities_p3[unique_entities_p3.type=='gene'].entity.to_list()]


    def create_prompt(self, prompt_config):

        prompt = "[BOS]"

        multi_modal_prefix = ''

        for k, v in prompt_config.items():
            if k=='instruction':
                prompt+=f'<{v}>' if isinstance(v, str) else "".join([f'<{v_i}>' for v_i in v])
            elif k=='up':
                if v:
                    prompt+=f'{multi_modal_prefix}<{k}>{v} </{k}>' if isinstance(v, str) else f'{multi_modal_prefix}<{k}>{" ".join(v)} </{k}>'
            elif k=='down':
                if v:
                    prompt+=f'{multi_modal_prefix}<{k}>{v} </{k}>' if isinstance(v, str) else f'{multi_modal_prefix}<{k}>{" ".join(v)} </{k}>'
            elif k=='age':
                if isinstance(v, int):
                    if prompt_config['species'].strip() == 'human':
                        prompt+=f'<{k}_individ>{v} </{k}_individ>'
                    elif prompt_config['species'].strip() == 'macaque':
                        prompt+=f'<{k}_individ>Macaca-{int(v/20)} </{k}_individ>'
            else:
                if v:
                    prompt+=f'<{k}>{v.strip()} </{k}>' if isinstance(v, str) else f'<{k}>{" ".join(v)} </{k}>'
                else:
                    prompt+=f'<{k}></{k}>'
        return prompt

    def custom_generate(self,
                        input_ids, 
                        device, 
                        max_new_tokens,
                        mode, 
                        temperature=0.8, 
                        top_p=0.2, top_k=3550, 
                        n_next_tokens=50, num_return_sequences=1, random_seed=137):

        torch.manual_seed(random_seed)

        # Set parameters
        # temperature - Higher value for more randomness, lower for more control
        # top_p - Probability threshold for nucleus sampling (aka top-p sampling)
        # top_k - Ignore logits below the top-k value to reduce randomness (if non-zero)
        # n_next_tokens - Number of top next tokens when predicting compounds

        # Generate sequences
        outputs = []
        next_token_compounds = [] 

        for _ in range(num_return_sequences):
            start_time = time.time()
            generated_sequence = []
            current_token = input_ids.clone()

            for _ in range(max_new_tokens):  # Maximum length of generated sequence
                # Forward pass through the model
                logits = self.model.forward(
                    input_ids=current_token
                )[0]

                # Apply temperature to logits
                if temperature != 1.0:
                    logits = logits / temperature

                # Apply top-p sampling (nucleus sampling)
                sorted_logits, sorted_indices = torch.sort(logits, descending=True)
                cumulative_probs = torch.cumsum(torch.softmax(sorted_logits, dim=-1), dim=-1)
                sorted_indices_to_remove = cumulative_probs > top_p

                if top_k > 0:
                    sorted_indices_to_remove[..., top_k:] = 1

                # Set the logit values of the removed indices to a very small negative value
                inf_tensor = torch.tensor(float("-inf")).type(torch.bfloat16).to(logits.device)

                logits = logits.where(sorted_indices_to_remove, inf_tensor)


                # Sample the next token
                if current_token[0][-1] == self.tokenizer.encode('<drug>')[0] and len(next_token_compounds)==0:
                    next_token_compounds.append(torch.topk(torch.softmax(logits, dim=-1)[0][len(current_token[0])-1, :].flatten(), n_next_tokens).indices)

                next_token = torch.multinomial(torch.softmax(logits, dim=-1)[0], num_samples=1)[len(current_token[0])-1, :].unsqueeze(0)


                # Append the sampled token to the generated sequence
                generated_sequence.append(next_token.item())

                # Stop generation if an end token is generated
                if next_token == self.tokenizer.eos_token_id:
                    break

                # Prepare input for the next iteration
                current_token = torch.cat((current_token, next_token), dim=-1)
            print(time.time()-start_time)
            outputs.append(generated_sequence)
        
        # Process generated up/down lists
        processed_outputs = {"up": [], "down": []}
        if mode in ['meta2diff', 'meta2diff2compound']:
            for output in outputs:
                up_split_index = output.index(self.tokenizer.convert_tokens_to_ids('</up>'))
                generated_up_raw = [i.strip() for i in self.tokenizer.convert_ids_to_tokens(output[:up_split_index])]
                generated_up = sorted(set(generated_up_raw) & set(self.unique_genes_p3), key = generated_up_raw.index)
                processed_outputs['up'].append(generated_up)
                
                down_split_index = output.index(self.tokenizer.convert_tokens_to_ids('</down>'))
                generated_down_raw = [i.strip() for i in self.tokenizer.convert_ids_to_tokens(output[up_split_index:down_split_index+1])]
                generated_down = sorted(set(generated_down_raw) & set(self.unique_genes_p3), key = generated_down_raw.index)
                processed_outputs['down'].append(generated_down)
                
        else:
            processed_outputs = outputs
        
        predicted_compounds_ids = [self.tokenizer.convert_ids_to_tokens(j) for j in next_token_compounds]
        predicted_compounds = []
        for j in predicted_compounds_ids:
            predicted_compounds.append([i.strip() for i in j])
        return processed_outputs, predicted_compounds, random_seed


    def __call__(self, data: Dict[str, Any]) -> Dict[str, str]:
        """
        Args:
            data (:dict:):
                The payload with the text prompt and generation parameters.
        """

        device = "cuda"
        parameters = data.pop("parameters", None)
        config_data = data.pop("inputs", None)
        mode = data.pop('mode', 'Not specified')
        
        prompt = self.create_prompt(config_data)
        
        inputs = self.tokenizer(prompt, return_tensors="pt")
        input_ids = inputs["input_ids"].to(device)

        max_new_tokens = self.model.config.max_seq_len - len(input_ids[0]) 
        try:
        
            generated_sequence, raw_next_token_generation, out_seed = self.custom_generate(input_ids = input_ids, 
                                                                                           max_new_tokens=max_new_tokens, mode=mode,
                                                                                           device=device, **parameters)
            next_token_generation = [sorted(set(i) & set(self.unique_compounds_p3), key = i.index) for i in raw_next_token_generation]

            if mode == "meta2diff":
                outputs = {"up": generated_sequence['up'], "down": generated_sequence['down']}
                out = {"output": outputs, "mode": mode, "message": "Done!", "input": prompt, 'random_seed': out_seed}
            elif mode == "meta2diff2compound":
                outputs = {"up": generated_sequence['up'], "down": generated_sequence['down']}
                out = {
                "output": outputs, "compounds": next_token_generation, "raw_output": raw_next_token_generation, "mode": mode, 
                    "message": "Done!", "input": prompt, 'random_seed': out_seed}
            elif mode == "diff2compound":
                outputs = generated_sequence
                out = {
                "output": outputs, "compounds": next_token_generation, "raw_output": raw_next_token_generation, "mode": mode, 
                    "message": "Done!", "input": prompt, 'random_seed': out_seed}
            else:
                out = {"message": f"Specify one of the following modes: meta2diff, meta2diff2compound, diff2compound. Your mode is: {mode}"}

        except Exception as e:
            print(e)
            outputs, next_token_generation = [None], [None]
            out = {"output": outputs, "mode": mode, 'message': f"{e}", "input": prompt, 'random_seed': 137}

        return out