BrightBlueCheese
commited on
Commit
•
7f92264
1
Parent(s):
90d0c74
app
Browse files- .ipynb_checkpoints/Untitled-checkpoint.ipynb +6 -0
- .ipynb_checkpoints/datamodule_finetune_sl-checkpoint.py +1 -1
- Untitled.ipynb +467 -0
- __pycache__/auto_evaluator_sl.cpython-311.pyc +0 -0
- __pycache__/chemllama_mtr.cpython-311.pyc +0 -0
- __pycache__/datamodule_finetune_sl.cpython-311.pyc +0 -0
- __pycache__/model_finetune_sl.cpython-311.pyc +0 -0
- __pycache__/tokenizer_sl.cpython-311.pyc +0 -0
- __pycache__/utils_sl.cpython-311.pyc +0 -0
- datamodule_finetune_sl.py +1 -1
.ipynb_checkpoints/Untitled-checkpoint.ipynb
ADDED
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"cells": [],
|
3 |
+
"metadata": {},
|
4 |
+
"nbformat": 4,
|
5 |
+
"nbformat_minor": 5
|
6 |
+
}
|
.ipynb_checkpoints/datamodule_finetune_sl-checkpoint.py
CHANGED
@@ -61,7 +61,7 @@ class CustomLlamaDatasetAbraham(Dataset):
|
|
61 |
return {
|
62 |
"input_ids": torch.tensor(local_encoded["input_ids"]),
|
63 |
"attention_mask": torch.tensor(local_encoded["attention_mask"]),
|
64 |
-
"labels":
|
65 |
}
|
66 |
|
67 |
class CustomFinetuneDataModule(L.LightningDataModule):
|
|
|
61 |
return {
|
62 |
"input_ids": torch.tensor(local_encoded["input_ids"]),
|
63 |
"attention_mask": torch.tensor(local_encoded["attention_mask"]),
|
64 |
+
"labels": torch.tensor(local_encoded["input_ids"]), # this one does not matter for sl
|
65 |
}
|
66 |
|
67 |
class CustomFinetuneDataModule(L.LightningDataModule):
|
Untitled.ipynb
ADDED
@@ -0,0 +1,467 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"cells": [
|
3 |
+
{
|
4 |
+
"cell_type": "code",
|
5 |
+
"execution_count": 1,
|
6 |
+
"id": "7e38e8a0-ff53-465c-9861-069d6dc54714",
|
7 |
+
"metadata": {},
|
8 |
+
"outputs": [],
|
9 |
+
"source": [
|
10 |
+
"import streamlit as st\n"
|
11 |
+
]
|
12 |
+
},
|
13 |
+
{
|
14 |
+
"cell_type": "code",
|
15 |
+
"execution_count": 2,
|
16 |
+
"id": "3c37a529-b0b4-4aed-a198-49f5e5bdbe02",
|
17 |
+
"metadata": {},
|
18 |
+
"outputs": [],
|
19 |
+
"source": [
|
20 |
+
"import sys\n",
|
21 |
+
"import os\n",
|
22 |
+
"import torch\n",
|
23 |
+
"from torch import nn\n",
|
24 |
+
"import torchmetrics\n",
|
25 |
+
"from transformers import LlamaModel, LlamaConfig\n",
|
26 |
+
"import numpy as np\n",
|
27 |
+
"import pandas as pd\n",
|
28 |
+
"import warnings\n",
|
29 |
+
"import lightning as L\n",
|
30 |
+
"torch.set_float32_matmul_precision('high')\n",
|
31 |
+
"warnings.filterwarnings(\"ignore\", module=\"pl_bolts\")"
|
32 |
+
]
|
33 |
+
},
|
34 |
+
{
|
35 |
+
"cell_type": "code",
|
36 |
+
"execution_count": 5,
|
37 |
+
"id": "1daba56d-a0e2-4be7-a2ea-52579726c201",
|
38 |
+
"metadata": {},
|
39 |
+
"outputs": [],
|
40 |
+
"source": [
|
41 |
+
"sys.path.append( '../')\n",
|
42 |
+
"\n",
|
43 |
+
"import tokenizer_sl, datamodule_finetune_sl, model_finetune_sl, chemllama_mtr, utils_sl\n",
|
44 |
+
"import auto_evaluator_sl\n",
|
45 |
+
"\n",
|
46 |
+
"from torch.utils.data import Dataset, DataLoader\n",
|
47 |
+
"from transformers import DataCollatorWithPadding\n",
|
48 |
+
"\n",
|
49 |
+
"torch.manual_seed(1004)\n",
|
50 |
+
"np.random.seed(1004)\n",
|
51 |
+
"\n",
|
52 |
+
"smiles_str = \"COO2\"\n",
|
53 |
+
"\n",
|
54 |
+
"solute_or_solvent = \"Solvent\"\n",
|
55 |
+
"\n"
|
56 |
+
]
|
57 |
+
},
|
58 |
+
{
|
59 |
+
"cell_type": "code",
|
60 |
+
"execution_count": 6,
|
61 |
+
"id": "7d3d996c-59b3-4079-83ef-818651add7ba",
|
62 |
+
"metadata": {},
|
63 |
+
"outputs": [],
|
64 |
+
"source": [
|
65 |
+
"class ChemLlama(nn.Module):\n",
|
66 |
+
" def __init__(\n",
|
67 |
+
" self,\n",
|
68 |
+
" max_position_embeddings=512,\n",
|
69 |
+
" vocab_size=591,\n",
|
70 |
+
" pad_token_id=0,\n",
|
71 |
+
" bos_token_id=12,\n",
|
72 |
+
" eos_token_id=13,\n",
|
73 |
+
" hidden_size=768,\n",
|
74 |
+
" intermediate_size=768,\n",
|
75 |
+
" num_labels=105,\n",
|
76 |
+
" attention_dropout=0.144,\n",
|
77 |
+
" num_hidden_layers=7,\n",
|
78 |
+
" num_attention_heads=8,\n",
|
79 |
+
" learning_rate=0.0001,\n",
|
80 |
+
" ):\n",
|
81 |
+
" super(ChemLlama, self).__init__()\n",
|
82 |
+
" \n",
|
83 |
+
" self.hidden_size = hidden_size\n",
|
84 |
+
" self.intermediate_size = intermediate_size\n",
|
85 |
+
" self.num_labels = num_labels\n",
|
86 |
+
" self.vocab_size = vocab_size\n",
|
87 |
+
" self.pad_token_id = pad_token_id\n",
|
88 |
+
" self.bos_token_id = bos_token_id\n",
|
89 |
+
" self.eos_token_id = eos_token_id\n",
|
90 |
+
" self.num_hidden_layers = num_hidden_layers\n",
|
91 |
+
" self.num_attention_heads = num_attention_heads\n",
|
92 |
+
" self.attention_dropout = attention_dropout\n",
|
93 |
+
" self.max_position_embeddings = max_position_embeddings\n",
|
94 |
+
"\n",
|
95 |
+
" self.mae = torchmetrics.MeanAbsoluteError()\n",
|
96 |
+
" self.mse = torchmetrics.MeanSquaredError()\n",
|
97 |
+
"\n",
|
98 |
+
" self.config_llama = LlamaConfig(\n",
|
99 |
+
" max_position_embeddings=self.max_position_embeddings,\n",
|
100 |
+
" vocab_size=self.vocab_size,\n",
|
101 |
+
" hidden_size=self.hidden_size,\n",
|
102 |
+
" intermediate_size=self.intermediate_size,\n",
|
103 |
+
" num_hidden_layers=self.num_hidden_layers,\n",
|
104 |
+
" num_attention_heads=self.num_attention_heads,\n",
|
105 |
+
" attention_dropout=self.attention_dropout,\n",
|
106 |
+
" pad_token_id=self.pad_token_id,\n",
|
107 |
+
" bos_token_id=self.bos_token_id,\n",
|
108 |
+
" eos_token_id=self.eos_token_id,\n",
|
109 |
+
" )\n",
|
110 |
+
"\n",
|
111 |
+
" self.loss_fn = nn.L1Loss()\n",
|
112 |
+
"\n",
|
113 |
+
" self.llama = LlamaModel(self.config_llama)\n",
|
114 |
+
" self.gelu = nn.GELU()\n",
|
115 |
+
" self.score = nn.Linear(self.hidden_size, self.num_labels)\n",
|
116 |
+
"\n",
|
117 |
+
" def forward(self, input_ids, attention_mask, labels=None):\n",
|
118 |
+
"\n",
|
119 |
+
" transformer_outputs = self.llama(\n",
|
120 |
+
" input_ids=input_ids, attention_mask=attention_mask\n",
|
121 |
+
" )\n",
|
122 |
+
"\n",
|
123 |
+
" hidden_states = transformer_outputs[0]\n",
|
124 |
+
" hidden_states = self.gelu(hidden_states)\n",
|
125 |
+
" logits = self.score(hidden_states)\n",
|
126 |
+
"\n",
|
127 |
+
" if input_ids is not None:\n",
|
128 |
+
" batch_size = input_ids.shape[0]\n",
|
129 |
+
" else:\n",
|
130 |
+
" batch_size = inputs_embeds.shape[0]\n",
|
131 |
+
"\n",
|
132 |
+
" if self.config_llama.pad_token_id is None and batch_size != 1:\n",
|
133 |
+
" raise ValueError(\n",
|
134 |
+
" \"Cannot handle batch sizes > 1 if no padding token is defined.\"\n",
|
135 |
+
" )\n",
|
136 |
+
" if self.config_llama.pad_token_id is None:\n",
|
137 |
+
" sequence_lengths = -1\n",
|
138 |
+
" else:\n",
|
139 |
+
" if input_ids is not None:\n",
|
140 |
+
" # if no pad token found, use modulo instead of reverse indexing for ONNX compatibility\n",
|
141 |
+
" sequence_lengths = (\n",
|
142 |
+
" torch.eq(input_ids, self.config_llama.pad_token_id).int().argmax(-1)\n",
|
143 |
+
" - 1\n",
|
144 |
+
" )\n",
|
145 |
+
" sequence_lengths = sequence_lengths % input_ids.shape[-1]\n",
|
146 |
+
" sequence_lengths = sequence_lengths.to(logits.device)\n",
|
147 |
+
" else:\n",
|
148 |
+
" sequence_lengths = -1\n",
|
149 |
+
" # raise ValueError(len(sequence_lengths), sequence_lengths)\n",
|
150 |
+
"\n",
|
151 |
+
" pooled_logits = logits[\n",
|
152 |
+
" torch.arange(batch_size, device=logits.device), sequence_lengths\n",
|
153 |
+
" ]\n",
|
154 |
+
" return pooled_logits\n",
|
155 |
+
"\n",
|
156 |
+
"\n",
|
157 |
+
"chemllama_mtr = ChemLlama()"
|
158 |
+
]
|
159 |
+
},
|
160 |
+
{
|
161 |
+
"cell_type": "code",
|
162 |
+
"execution_count": 7,
|
163 |
+
"id": "da586e81-ace8-489d-a11a-ae44a0ed2369",
|
164 |
+
"metadata": {},
|
165 |
+
"outputs": [
|
166 |
+
{
|
167 |
+
"name": "stdout",
|
168 |
+
"output_type": "stream",
|
169 |
+
"text": [
|
170 |
+
"llama.embed_tokens.weight False\n",
|
171 |
+
"llama.layers.0.self_attn.q_proj.weight False\n",
|
172 |
+
"llama.layers.0.self_attn.k_proj.weight False\n",
|
173 |
+
"llama.layers.0.self_attn.v_proj.weight False\n",
|
174 |
+
"llama.layers.0.self_attn.o_proj.weight False\n",
|
175 |
+
"llama.layers.0.mlp.gate_proj.weight False\n",
|
176 |
+
"llama.layers.0.mlp.up_proj.weight False\n",
|
177 |
+
"llama.layers.0.mlp.down_proj.weight False\n",
|
178 |
+
"llama.layers.0.input_layernorm.weight False\n",
|
179 |
+
"llama.layers.0.post_attention_layernorm.weight False\n",
|
180 |
+
"llama.layers.1.self_attn.q_proj.weight False\n",
|
181 |
+
"llama.layers.1.self_attn.k_proj.weight False\n",
|
182 |
+
"llama.layers.1.self_attn.v_proj.weight False\n",
|
183 |
+
"llama.layers.1.self_attn.o_proj.weight False\n",
|
184 |
+
"llama.layers.1.mlp.gate_proj.weight False\n",
|
185 |
+
"llama.layers.1.mlp.up_proj.weight False\n",
|
186 |
+
"llama.layers.1.mlp.down_proj.weight False\n",
|
187 |
+
"llama.layers.1.input_layernorm.weight False\n",
|
188 |
+
"llama.layers.1.post_attention_layernorm.weight False\n",
|
189 |
+
"llama.layers.2.self_attn.q_proj.weight False\n",
|
190 |
+
"llama.layers.2.self_attn.k_proj.weight False\n",
|
191 |
+
"llama.layers.2.self_attn.v_proj.weight False\n",
|
192 |
+
"llama.layers.2.self_attn.o_proj.weight False\n",
|
193 |
+
"llama.layers.2.mlp.gate_proj.weight False\n",
|
194 |
+
"llama.layers.2.mlp.up_proj.weight False\n",
|
195 |
+
"llama.layers.2.mlp.down_proj.weight False\n",
|
196 |
+
"llama.layers.2.input_layernorm.weight False\n",
|
197 |
+
"llama.layers.2.post_attention_layernorm.weight False\n",
|
198 |
+
"llama.layers.3.self_attn.q_proj.weight False\n",
|
199 |
+
"llama.layers.3.self_attn.k_proj.weight False\n",
|
200 |
+
"llama.layers.3.self_attn.v_proj.weight False\n",
|
201 |
+
"llama.layers.3.self_attn.o_proj.weight False\n",
|
202 |
+
"llama.layers.3.mlp.gate_proj.weight False\n",
|
203 |
+
"llama.layers.3.mlp.up_proj.weight False\n",
|
204 |
+
"llama.layers.3.mlp.down_proj.weight False\n",
|
205 |
+
"llama.layers.3.input_layernorm.weight False\n",
|
206 |
+
"llama.layers.3.post_attention_layernorm.weight False\n",
|
207 |
+
"llama.layers.4.self_attn.q_proj.weight False\n",
|
208 |
+
"llama.layers.4.self_attn.k_proj.weight False\n",
|
209 |
+
"llama.layers.4.self_attn.v_proj.weight False\n",
|
210 |
+
"llama.layers.4.self_attn.o_proj.weight False\n",
|
211 |
+
"llama.layers.4.mlp.gate_proj.weight False\n",
|
212 |
+
"llama.layers.4.mlp.up_proj.weight False\n",
|
213 |
+
"llama.layers.4.mlp.down_proj.weight False\n",
|
214 |
+
"llama.layers.4.input_layernorm.weight False\n",
|
215 |
+
"llama.layers.4.post_attention_layernorm.weight False\n",
|
216 |
+
"llama.layers.5.self_attn.q_proj.weight False\n",
|
217 |
+
"llama.layers.5.self_attn.k_proj.weight False\n",
|
218 |
+
"llama.layers.5.self_attn.v_proj.weight False\n",
|
219 |
+
"llama.layers.5.self_attn.o_proj.weight False\n",
|
220 |
+
"llama.layers.5.mlp.gate_proj.weight False\n",
|
221 |
+
"llama.layers.5.mlp.up_proj.weight False\n",
|
222 |
+
"llama.layers.5.mlp.down_proj.weight False\n",
|
223 |
+
"llama.layers.5.input_layernorm.weight False\n",
|
224 |
+
"llama.layers.5.post_attention_layernorm.weight False\n",
|
225 |
+
"llama.layers.6.self_attn.q_proj.weight False\n",
|
226 |
+
"llama.layers.6.self_attn.k_proj.weight False\n",
|
227 |
+
"llama.layers.6.self_attn.v_proj.weight False\n",
|
228 |
+
"llama.layers.6.self_attn.o_proj.weight False\n",
|
229 |
+
"llama.layers.6.mlp.gate_proj.weight False\n",
|
230 |
+
"llama.layers.6.mlp.up_proj.weight False\n",
|
231 |
+
"llama.layers.6.mlp.down_proj.weight False\n",
|
232 |
+
"llama.layers.6.input_layernorm.weight False\n",
|
233 |
+
"llama.layers.6.post_attention_layernorm.weight False\n",
|
234 |
+
"llama.norm.weight False\n",
|
235 |
+
"score.weight False\n",
|
236 |
+
"score.bias False\n"
|
237 |
+
]
|
238 |
+
}
|
239 |
+
],
|
240 |
+
"source": [
|
241 |
+
"class ChemLlama_FT(nn.Module):\n",
|
242 |
+
" def __init__(\n",
|
243 |
+
" self,\n",
|
244 |
+
" model_mtr,\n",
|
245 |
+
" linear_param:int=64,\n",
|
246 |
+
" use_freeze:bool=True,\n",
|
247 |
+
" *args, **kwargs\n",
|
248 |
+
" ):\n",
|
249 |
+
" super(ChemLlama_FT, self).__init__()\n",
|
250 |
+
" # self.save_hyperparameters()\n",
|
251 |
+
"\n",
|
252 |
+
" self.model_mtr = model_mtr\n",
|
253 |
+
" if use_freeze:\n",
|
254 |
+
" # self.model_mtr.freeze()\n",
|
255 |
+
" for name, param in model_mtr.named_parameters():\n",
|
256 |
+
" param.requires_grad = False\n",
|
257 |
+
" print(name, param.requires_grad)\n",
|
258 |
+
" \n",
|
259 |
+
" self.gelu = nn.GELU()\n",
|
260 |
+
" self.linear1 = nn.Linear(self.model_mtr.num_labels, linear_param)\n",
|
261 |
+
" self.linear2 = nn.Linear(linear_param, linear_param)\n",
|
262 |
+
" self.regression = nn.Linear(linear_param, 5)\n",
|
263 |
+
"\n",
|
264 |
+
" self.loss_fn = nn.L1Loss()\n",
|
265 |
+
"\n",
|
266 |
+
" def forward(self, input_ids, attention_mask, labels=None):\n",
|
267 |
+
" x = self.model_mtr(input_ids=input_ids, attention_mask=attention_mask)\n",
|
268 |
+
" x = self.gelu(x)\n",
|
269 |
+
" x = self.linear1(x)\n",
|
270 |
+
" x = self.gelu(x)\n",
|
271 |
+
" x = self.linear2(x)\n",
|
272 |
+
" x = self.gelu(x)\n",
|
273 |
+
" x = self.regression(x)\n",
|
274 |
+
" \n",
|
275 |
+
" return x\n",
|
276 |
+
" \n",
|
277 |
+
"chemllama_ft = ChemLlama_FT(model_mtr=chemllama_mtr)"
|
278 |
+
]
|
279 |
+
},
|
280 |
+
{
|
281 |
+
"cell_type": "code",
|
282 |
+
"execution_count": 9,
|
283 |
+
"id": "49537588-bad0-44ff-b7fd-73683cdb2f6c",
|
284 |
+
"metadata": {},
|
285 |
+
"outputs": [],
|
286 |
+
"source": [
|
287 |
+
"# I just reused our previous research code with some modifications.\n",
|
288 |
+
"dir_main = \"../\"\n",
|
289 |
+
"\n",
|
290 |
+
"max_seq_length = 512\n",
|
291 |
+
"\n",
|
292 |
+
"tokenizer = tokenizer_sl.fn_load_tokenizer_llama(\n",
|
293 |
+
" max_seq_length=max_seq_length,\n",
|
294 |
+
")\n",
|
295 |
+
"max_length = max_seq_length\n",
|
296 |
+
"num_workers = 2\n",
|
297 |
+
"\n",
|
298 |
+
"## FT\n",
|
299 |
+
"\n",
|
300 |
+
"dir_model_ft_to_save = f\"{dir_main}/SolLlama-mtr\"\n",
|
301 |
+
"# name_model_ft = 'Solvent.pt'\n",
|
302 |
+
"name_model_ft = f\"{solute_or_solvent}.pt\""
|
303 |
+
]
|
304 |
+
},
|
305 |
+
{
|
306 |
+
"cell_type": "code",
|
307 |
+
"execution_count": 20,
|
308 |
+
"id": "cc155008-a7f1-4dd1-8fc3-ad299a5938a6",
|
309 |
+
"metadata": {},
|
310 |
+
"outputs": [],
|
311 |
+
"source": [
|
312 |
+
"device = 'cpu'\n",
|
313 |
+
"# Predict\n",
|
314 |
+
"local_model_ft = utils_sl.load_model_ft_with(\n",
|
315 |
+
" class_model_ft=chemllama_ft, \n",
|
316 |
+
" dir_model_ft=dir_model_ft_to_save,\n",
|
317 |
+
" name_model_ft=name_model_ft\n",
|
318 |
+
").to(device)\n",
|
319 |
+
"\n",
|
320 |
+
"# result = trainer.predict(local_model_ft, data_module)\n",
|
321 |
+
"# result_pred = list()\n",
|
322 |
+
"# result_label = list()\n",
|
323 |
+
"# for bat in range(len(result)):\n",
|
324 |
+
"# result_pred.append(result[bat][0].squeeze())\n",
|
325 |
+
"# result_label.append(result[bat][1])\n",
|
326 |
+
"\n",
|
327 |
+
"# with open('./smiles_str.txt', 'r') as file:\n",
|
328 |
+
"# smiles_str = file.readline()\n",
|
329 |
+
" \n",
|
330 |
+
"dataset_test = datamodule_finetune_sl.CustomLlamaDatasetAbraham(\n",
|
331 |
+
" df=pd.DataFrame([smiles_str]),\n",
|
332 |
+
" tokenizer=tokenizer,\n",
|
333 |
+
" max_seq_length=max_length\n",
|
334 |
+
")\n",
|
335 |
+
"\n",
|
336 |
+
"data_collator = DataCollatorWithPadding(tokenizer)\n",
|
337 |
+
"dataloader_test = DataLoader(dataset_test, batch_size=1, shuffle=False, collate_fn=data_collator)"
|
338 |
+
]
|
339 |
+
},
|
340 |
+
{
|
341 |
+
"cell_type": "code",
|
342 |
+
"execution_count": 31,
|
343 |
+
"id": "69baeffd-a2cb-439c-be46-69ee4fc5fea1",
|
344 |
+
"metadata": {},
|
345 |
+
"outputs": [
|
346 |
+
{
|
347 |
+
"data": {
|
348 |
+
"text/plain": [
|
349 |
+
"0 COO2\n",
|
350 |
+
"Name: 0, dtype: object"
|
351 |
+
]
|
352 |
+
},
|
353 |
+
"execution_count": 31,
|
354 |
+
"metadata": {},
|
355 |
+
"output_type": "execute_result"
|
356 |
+
}
|
357 |
+
],
|
358 |
+
"source": [
|
359 |
+
"pd.DataFrame([smiles_str]).iloc[:,0:].iloc[0]"
|
360 |
+
]
|
361 |
+
},
|
362 |
+
{
|
363 |
+
"cell_type": "code",
|
364 |
+
"execution_count": 22,
|
365 |
+
"id": "7994f626-ca68-4ef1-811d-c2b684cd62ce",
|
366 |
+
"metadata": {},
|
367 |
+
"outputs": [
|
368 |
+
{
|
369 |
+
"data": {
|
370 |
+
"text/plain": [
|
371 |
+
"<datamodule_finetune_sl.CustomLlamaDatasetAbraham at 0x7f81a4f6cf10>"
|
372 |
+
]
|
373 |
+
},
|
374 |
+
"execution_count": 22,
|
375 |
+
"metadata": {},
|
376 |
+
"output_type": "execute_result"
|
377 |
+
}
|
378 |
+
],
|
379 |
+
"source": [
|
380 |
+
"dataset_test"
|
381 |
+
]
|
382 |
+
},
|
383 |
+
{
|
384 |
+
"cell_type": "code",
|
385 |
+
"execution_count": null,
|
386 |
+
"id": "3b173642-1695-40dc-82b5-0e7b775fff38",
|
387 |
+
"metadata": {},
|
388 |
+
"outputs": [],
|
389 |
+
"source": [
|
390 |
+
"data_loader_valid = DataLoader(dataset_valid, batch_size=int(batch_size*1.5), shuffle=False, collate_fn=data_collator, num_workers=4, pin_memory=True)"
|
391 |
+
]
|
392 |
+
},
|
393 |
+
{
|
394 |
+
"cell_type": "code",
|
395 |
+
"execution_count": 21,
|
396 |
+
"id": "a6d6145b-d5f9-44e4-85ca-f27b8c8a339d",
|
397 |
+
"metadata": {},
|
398 |
+
"outputs": [
|
399 |
+
{
|
400 |
+
"ename": "ValueError",
|
401 |
+
"evalue": "Unable to create tensor, you should probably activate truncation and/or padding with 'padding=True' 'truncation=True' to have batched tensors with the same length. Perhaps your features (`labels` in this case) have excessive nesting (inputs type `list` where type `int` is expected).",
|
402 |
+
"output_type": "error",
|
403 |
+
"traceback": [
|
404 |
+
"\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
|
405 |
+
"\u001b[0;31mRuntimeError\u001b[0m Traceback (most recent call last)",
|
406 |
+
"File \u001b[0;32m~/chemllm/lib/python3.11/site-packages/transformers/tokenization_utils_base.py:759\u001b[0m, in \u001b[0;36mBatchEncoding.convert_to_tensors\u001b[0;34m(self, tensor_type, prepend_batch_axis)\u001b[0m\n\u001b[1;32m 758\u001b[0m \u001b[38;5;28;01mif\u001b[39;00m \u001b[38;5;129;01mnot\u001b[39;00m is_tensor(value):\n\u001b[0;32m--> 759\u001b[0m tensor \u001b[38;5;241m=\u001b[39m \u001b[43mas_tensor\u001b[49m\u001b[43m(\u001b[49m\u001b[43mvalue\u001b[49m\u001b[43m)\u001b[49m\n\u001b[1;32m 761\u001b[0m \u001b[38;5;66;03m# Removing this for now in favor of controlling the shape with `prepend_batch_axis`\u001b[39;00m\n\u001b[1;32m 762\u001b[0m \u001b[38;5;66;03m# # at-least2d\u001b[39;00m\n\u001b[1;32m 763\u001b[0m \u001b[38;5;66;03m# if tensor.ndim > 2:\u001b[39;00m\n\u001b[1;32m 764\u001b[0m \u001b[38;5;66;03m# tensor = tensor.squeeze(0)\u001b[39;00m\n\u001b[1;32m 765\u001b[0m \u001b[38;5;66;03m# elif tensor.ndim < 2:\u001b[39;00m\n\u001b[1;32m 766\u001b[0m \u001b[38;5;66;03m# tensor = tensor[None, :]\u001b[39;00m\n",
|
407 |
+
"File \u001b[0;32m~/chemllm/lib/python3.11/site-packages/transformers/tokenization_utils_base.py:721\u001b[0m, in \u001b[0;36mBatchEncoding.convert_to_tensors.<locals>.as_tensor\u001b[0;34m(value, dtype)\u001b[0m\n\u001b[1;32m 720\u001b[0m \u001b[38;5;28;01mreturn\u001b[39;00m torch\u001b[38;5;241m.\u001b[39mtensor(np\u001b[38;5;241m.\u001b[39marray(value))\n\u001b[0;32m--> 721\u001b[0m \u001b[38;5;28;01mreturn\u001b[39;00m \u001b[43mtorch\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mtensor\u001b[49m\u001b[43m(\u001b[49m\u001b[43mvalue\u001b[49m\u001b[43m)\u001b[49m\n",
|
408 |
+
"\u001b[0;31mRuntimeError\u001b[0m: Could not infer dtype of NoneType",
|
409 |
+
"\nThe above exception was the direct cause of the following exception:\n",
|
410 |
+
"\u001b[0;31mValueError\u001b[0m Traceback (most recent call last)",
|
411 |
+
"Cell \u001b[0;32mIn[21], line 4\u001b[0m\n\u001b[1;32m 2\u001b[0m local_model_ft\u001b[38;5;241m.\u001b[39meval()\n\u001b[1;32m 3\u001b[0m \u001b[38;5;28;01mwith\u001b[39;00m torch\u001b[38;5;241m.\u001b[39minference_mode():\n\u001b[0;32m----> 4\u001b[0m \u001b[38;5;28;01mfor\u001b[39;00m i, v_batch \u001b[38;5;129;01min\u001b[39;00m \u001b[38;5;28menumerate\u001b[39m(dataloader_test):\n\u001b[1;32m 5\u001b[0m \u001b[38;5;28;01mbreak\u001b[39;00m\n\u001b[1;32m 6\u001b[0m \u001b[38;5;66;03m# v_input_ids = v_batch['input_ids'].to(device)\u001b[39;00m\n\u001b[1;32m 7\u001b[0m \u001b[38;5;66;03m# v_attention_mask = v_batch['attention_mask'].to(device)\u001b[39;00m\n\u001b[1;32m 8\u001b[0m \u001b[38;5;66;03m# # v_y_labels = v_batch['labels'].to(device)\u001b[39;00m\n\u001b[1;32m 9\u001b[0m \u001b[38;5;66;03m# v_y_logits = local_model_ft(input_ids=v_input_ids, attention_mask=v_attention_mask)\u001b[39;00m\n\u001b[1;32m 10\u001b[0m \u001b[38;5;66;03m# list_predictions.append(v_y_logits[0][0].tolist())\u001b[39;00m\n",
|
412 |
+
"File \u001b[0;32m~/chemllm/lib/python3.11/site-packages/torch/utils/data/dataloader.py:634\u001b[0m, in \u001b[0;36m_BaseDataLoaderIter.__next__\u001b[0;34m(self)\u001b[0m\n\u001b[1;32m 631\u001b[0m \u001b[38;5;28;01mif\u001b[39;00m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_sampler_iter \u001b[38;5;129;01mis\u001b[39;00m \u001b[38;5;28;01mNone\u001b[39;00m:\n\u001b[1;32m 632\u001b[0m \u001b[38;5;66;03m# TODO(https://github.com/pytorch/pytorch/issues/76750)\u001b[39;00m\n\u001b[1;32m 633\u001b[0m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_reset() \u001b[38;5;66;03m# type: ignore[call-arg]\u001b[39;00m\n\u001b[0;32m--> 634\u001b[0m data \u001b[38;5;241m=\u001b[39m \u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43m_next_data\u001b[49m\u001b[43m(\u001b[49m\u001b[43m)\u001b[49m\n\u001b[1;32m 635\u001b[0m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_num_yielded \u001b[38;5;241m+\u001b[39m\u001b[38;5;241m=\u001b[39m \u001b[38;5;241m1\u001b[39m\n\u001b[1;32m 636\u001b[0m \u001b[38;5;28;01mif\u001b[39;00m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_dataset_kind \u001b[38;5;241m==\u001b[39m _DatasetKind\u001b[38;5;241m.\u001b[39mIterable \u001b[38;5;129;01mand\u001b[39;00m \\\n\u001b[1;32m 637\u001b[0m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_IterableDataset_len_called \u001b[38;5;129;01mis\u001b[39;00m \u001b[38;5;129;01mnot\u001b[39;00m \u001b[38;5;28;01mNone\u001b[39;00m \u001b[38;5;129;01mand\u001b[39;00m \\\n\u001b[1;32m 638\u001b[0m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_num_yielded \u001b[38;5;241m>\u001b[39m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_IterableDataset_len_called:\n",
|
413 |
+
"File \u001b[0;32m~/chemllm/lib/python3.11/site-packages/torch/utils/data/dataloader.py:678\u001b[0m, in \u001b[0;36m_SingleProcessDataLoaderIter._next_data\u001b[0;34m(self)\u001b[0m\n\u001b[1;32m 676\u001b[0m \u001b[38;5;28;01mdef\u001b[39;00m \u001b[38;5;21m_next_data\u001b[39m(\u001b[38;5;28mself\u001b[39m):\n\u001b[1;32m 677\u001b[0m index \u001b[38;5;241m=\u001b[39m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_next_index() \u001b[38;5;66;03m# may raise StopIteration\u001b[39;00m\n\u001b[0;32m--> 678\u001b[0m data \u001b[38;5;241m=\u001b[39m \u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43m_dataset_fetcher\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mfetch\u001b[49m\u001b[43m(\u001b[49m\u001b[43mindex\u001b[49m\u001b[43m)\u001b[49m \u001b[38;5;66;03m# may raise StopIteration\u001b[39;00m\n\u001b[1;32m 679\u001b[0m \u001b[38;5;28;01mif\u001b[39;00m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_pin_memory:\n\u001b[1;32m 680\u001b[0m data \u001b[38;5;241m=\u001b[39m _utils\u001b[38;5;241m.\u001b[39mpin_memory\u001b[38;5;241m.\u001b[39mpin_memory(data, \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_pin_memory_device)\n",
|
414 |
+
"File \u001b[0;32m~/chemllm/lib/python3.11/site-packages/torch/utils/data/_utils/fetch.py:54\u001b[0m, in \u001b[0;36m_MapDatasetFetcher.fetch\u001b[0;34m(self, possibly_batched_index)\u001b[0m\n\u001b[1;32m 52\u001b[0m \u001b[38;5;28;01melse\u001b[39;00m:\n\u001b[1;32m 53\u001b[0m data \u001b[38;5;241m=\u001b[39m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39mdataset[possibly_batched_index]\n\u001b[0;32m---> 54\u001b[0m \u001b[38;5;28;01mreturn\u001b[39;00m \u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mcollate_fn\u001b[49m\u001b[43m(\u001b[49m\u001b[43mdata\u001b[49m\u001b[43m)\u001b[49m\n",
|
415 |
+
"File \u001b[0;32m~/chemllm/lib/python3.11/site-packages/transformers/data/data_collator.py:271\u001b[0m, in \u001b[0;36mDataCollatorWithPadding.__call__\u001b[0;34m(self, features)\u001b[0m\n\u001b[1;32m 270\u001b[0m \u001b[38;5;28;01mdef\u001b[39;00m \u001b[38;5;21m__call__\u001b[39m(\u001b[38;5;28mself\u001b[39m, features: List[Dict[\u001b[38;5;28mstr\u001b[39m, Any]]) \u001b[38;5;241m-\u001b[39m\u001b[38;5;241m>\u001b[39m Dict[\u001b[38;5;28mstr\u001b[39m, Any]:\n\u001b[0;32m--> 271\u001b[0m batch \u001b[38;5;241m=\u001b[39m \u001b[43mpad_without_fast_tokenizer_warning\u001b[49m\u001b[43m(\u001b[49m\n\u001b[1;32m 272\u001b[0m \u001b[43m \u001b[49m\u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mtokenizer\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m 273\u001b[0m \u001b[43m \u001b[49m\u001b[43mfeatures\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m 274\u001b[0m \u001b[43m \u001b[49m\u001b[43mpadding\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mpadding\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m 275\u001b[0m \u001b[43m \u001b[49m\u001b[43mmax_length\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mmax_length\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m 276\u001b[0m \u001b[43m \u001b[49m\u001b[43mpad_to_multiple_of\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mpad_to_multiple_of\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m 277\u001b[0m \u001b[43m \u001b[49m\u001b[43mreturn_tensors\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mreturn_tensors\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m 278\u001b[0m \u001b[43m \u001b[49m\u001b[43m)\u001b[49m\n\u001b[1;32m 279\u001b[0m \u001b[38;5;28;01mif\u001b[39;00m \u001b[38;5;124m\"\u001b[39m\u001b[38;5;124mlabel\u001b[39m\u001b[38;5;124m\"\u001b[39m \u001b[38;5;129;01min\u001b[39;00m batch:\n\u001b[1;32m 280\u001b[0m batch[\u001b[38;5;124m\"\u001b[39m\u001b[38;5;124mlabels\u001b[39m\u001b[38;5;124m\"\u001b[39m] \u001b[38;5;241m=\u001b[39m batch[\u001b[38;5;124m\"\u001b[39m\u001b[38;5;124mlabel\u001b[39m\u001b[38;5;124m\"\u001b[39m]\n",
|
416 |
+
"File \u001b[0;32m~/chemllm/lib/python3.11/site-packages/transformers/data/data_collator.py:66\u001b[0m, in \u001b[0;36mpad_without_fast_tokenizer_warning\u001b[0;34m(tokenizer, *pad_args, **pad_kwargs)\u001b[0m\n\u001b[1;32m 63\u001b[0m tokenizer\u001b[38;5;241m.\u001b[39mdeprecation_warnings[\u001b[38;5;124m\"\u001b[39m\u001b[38;5;124mAsking-to-pad-a-fast-tokenizer\u001b[39m\u001b[38;5;124m\"\u001b[39m] \u001b[38;5;241m=\u001b[39m \u001b[38;5;28;01mTrue\u001b[39;00m\n\u001b[1;32m 65\u001b[0m \u001b[38;5;28;01mtry\u001b[39;00m:\n\u001b[0;32m---> 66\u001b[0m padded \u001b[38;5;241m=\u001b[39m \u001b[43mtokenizer\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mpad\u001b[49m\u001b[43m(\u001b[49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[43mpad_args\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[43mpad_kwargs\u001b[49m\u001b[43m)\u001b[49m\n\u001b[1;32m 67\u001b[0m \u001b[38;5;28;01mfinally\u001b[39;00m:\n\u001b[1;32m 68\u001b[0m \u001b[38;5;66;03m# Restore the state of the warning.\u001b[39;00m\n\u001b[1;32m 69\u001b[0m tokenizer\u001b[38;5;241m.\u001b[39mdeprecation_warnings[\u001b[38;5;124m\"\u001b[39m\u001b[38;5;124mAsking-to-pad-a-fast-tokenizer\u001b[39m\u001b[38;5;124m\"\u001b[39m] \u001b[38;5;241m=\u001b[39m warning_state\n",
|
417 |
+
"File \u001b[0;32m~/chemllm/lib/python3.11/site-packages/transformers/tokenization_utils_base.py:3369\u001b[0m, in \u001b[0;36mPreTrainedTokenizerBase.pad\u001b[0;34m(self, encoded_inputs, padding, max_length, pad_to_multiple_of, return_attention_mask, return_tensors, verbose)\u001b[0m\n\u001b[1;32m 3366\u001b[0m batch_outputs[key] \u001b[38;5;241m=\u001b[39m []\n\u001b[1;32m 3367\u001b[0m batch_outputs[key]\u001b[38;5;241m.\u001b[39mappend(value)\n\u001b[0;32m-> 3369\u001b[0m \u001b[38;5;28;01mreturn\u001b[39;00m \u001b[43mBatchEncoding\u001b[49m\u001b[43m(\u001b[49m\u001b[43mbatch_outputs\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mtensor_type\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43mreturn_tensors\u001b[49m\u001b[43m)\u001b[49m\n",
|
418 |
+
"File \u001b[0;32m~/chemllm/lib/python3.11/site-packages/transformers/tokenization_utils_base.py:224\u001b[0m, in \u001b[0;36mBatchEncoding.__init__\u001b[0;34m(self, data, encoding, tensor_type, prepend_batch_axis, n_sequences)\u001b[0m\n\u001b[1;32m 220\u001b[0m n_sequences \u001b[38;5;241m=\u001b[39m encoding[\u001b[38;5;241m0\u001b[39m]\u001b[38;5;241m.\u001b[39mn_sequences\n\u001b[1;32m 222\u001b[0m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_n_sequences \u001b[38;5;241m=\u001b[39m n_sequences\n\u001b[0;32m--> 224\u001b[0m \u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mconvert_to_tensors\u001b[49m\u001b[43m(\u001b[49m\u001b[43mtensor_type\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43mtensor_type\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mprepend_batch_axis\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43mprepend_batch_axis\u001b[49m\u001b[43m)\u001b[49m\n",
|
419 |
+
"File \u001b[0;32m~/chemllm/lib/python3.11/site-packages/transformers/tokenization_utils_base.py:775\u001b[0m, in \u001b[0;36mBatchEncoding.convert_to_tensors\u001b[0;34m(self, tensor_type, prepend_batch_axis)\u001b[0m\n\u001b[1;32m 770\u001b[0m \u001b[38;5;28;01mif\u001b[39;00m key \u001b[38;5;241m==\u001b[39m \u001b[38;5;124m\"\u001b[39m\u001b[38;5;124moverflowing_tokens\u001b[39m\u001b[38;5;124m\"\u001b[39m:\n\u001b[1;32m 771\u001b[0m \u001b[38;5;28;01mraise\u001b[39;00m \u001b[38;5;167;01mValueError\u001b[39;00m(\n\u001b[1;32m 772\u001b[0m \u001b[38;5;124m\"\u001b[39m\u001b[38;5;124mUnable to create tensor returning overflowing tokens of different lengths. \u001b[39m\u001b[38;5;124m\"\u001b[39m\n\u001b[1;32m 773\u001b[0m \u001b[38;5;124m\"\u001b[39m\u001b[38;5;124mPlease see if a fast version of this tokenizer is available to have this feature available.\u001b[39m\u001b[38;5;124m\"\u001b[39m\n\u001b[1;32m 774\u001b[0m ) \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01me\u001b[39;00m\n\u001b[0;32m--> 775\u001b[0m \u001b[38;5;28;01mraise\u001b[39;00m \u001b[38;5;167;01mValueError\u001b[39;00m(\n\u001b[1;32m 776\u001b[0m \u001b[38;5;124m\"\u001b[39m\u001b[38;5;124mUnable to create tensor, you should probably activate truncation and/or padding with\u001b[39m\u001b[38;5;124m\"\u001b[39m\n\u001b[1;32m 777\u001b[0m \u001b[38;5;124m\"\u001b[39m\u001b[38;5;124m \u001b[39m\u001b[38;5;124m'\u001b[39m\u001b[38;5;124mpadding=True\u001b[39m\u001b[38;5;124m'\u001b[39m\u001b[38;5;124m \u001b[39m\u001b[38;5;124m'\u001b[39m\u001b[38;5;124mtruncation=True\u001b[39m\u001b[38;5;124m'\u001b[39m\u001b[38;5;124m to have batched tensors with the same length. Perhaps your\u001b[39m\u001b[38;5;124m\"\u001b[39m\n\u001b[1;32m 778\u001b[0m \u001b[38;5;124mf\u001b[39m\u001b[38;5;124m\"\u001b[39m\u001b[38;5;124m features (`\u001b[39m\u001b[38;5;132;01m{\u001b[39;00mkey\u001b[38;5;132;01m}\u001b[39;00m\u001b[38;5;124m` in this case) have excessive nesting (inputs type `list` where type `int` is\u001b[39m\u001b[38;5;124m\"\u001b[39m\n\u001b[1;32m 779\u001b[0m \u001b[38;5;124m\"\u001b[39m\u001b[38;5;124m expected).\u001b[39m\u001b[38;5;124m\"\u001b[39m\n\u001b[1;32m 780\u001b[0m ) \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01me\u001b[39;00m\n\u001b[1;32m 782\u001b[0m \u001b[38;5;28;01mreturn\u001b[39;00m \u001b[38;5;28mself\u001b[39m\n",
|
420 |
+
"\u001b[0;31mValueError\u001b[0m: Unable to create tensor, you should probably activate truncation and/or padding with 'padding=True' 'truncation=True' to have batched tensors with the same length. Perhaps your features (`labels` in this case) have excessive nesting (inputs type `list` where type `int` is expected)."
|
421 |
+
]
|
422 |
+
}
|
423 |
+
],
|
424 |
+
"source": [
|
425 |
+
"list_predictions = []\n",
|
426 |
+
"local_model_ft.eval()\n",
|
427 |
+
"with torch.inference_mode():\n",
|
428 |
+
" for i, v_batch in enumerate(dataloader_test):\n",
|
429 |
+
" break\n",
|
430 |
+
" # v_input_ids = v_batch['input_ids'].to(device)\n",
|
431 |
+
" # v_attention_mask = v_batch['attention_mask'].to(device)\n",
|
432 |
+
" # # v_y_labels = v_batch['labels'].to(device)\n",
|
433 |
+
" # v_y_logits = local_model_ft(input_ids=v_input_ids, attention_mask=v_attention_mask)\n",
|
434 |
+
" # list_predictions.append(v_y_logits[0][0].tolist())"
|
435 |
+
]
|
436 |
+
},
|
437 |
+
{
|
438 |
+
"cell_type": "code",
|
439 |
+
"execution_count": null,
|
440 |
+
"id": "7bc3e296-6871-45fb-8459-78eadc36bb61",
|
441 |
+
"metadata": {},
|
442 |
+
"outputs": [],
|
443 |
+
"source": []
|
444 |
+
}
|
445 |
+
],
|
446 |
+
"metadata": {
|
447 |
+
"kernelspec": {
|
448 |
+
"display_name": "chemllm",
|
449 |
+
"language": "python",
|
450 |
+
"name": "chemllm"
|
451 |
+
},
|
452 |
+
"language_info": {
|
453 |
+
"codemirror_mode": {
|
454 |
+
"name": "ipython",
|
455 |
+
"version": 3
|
456 |
+
},
|
457 |
+
"file_extension": ".py",
|
458 |
+
"mimetype": "text/x-python",
|
459 |
+
"name": "python",
|
460 |
+
"nbconvert_exporter": "python",
|
461 |
+
"pygments_lexer": "ipython3",
|
462 |
+
"version": "3.11.3"
|
463 |
+
}
|
464 |
+
},
|
465 |
+
"nbformat": 4,
|
466 |
+
"nbformat_minor": 5
|
467 |
+
}
|
__pycache__/auto_evaluator_sl.cpython-311.pyc
ADDED
Binary file (9.12 kB). View file
|
|
__pycache__/chemllama_mtr.cpython-311.pyc
ADDED
Binary file (7.89 kB). View file
|
|
__pycache__/datamodule_finetune_sl.cpython-311.pyc
ADDED
Binary file (5.45 kB). View file
|
|
__pycache__/model_finetune_sl.cpython-311.pyc
ADDED
Binary file (6.43 kB). View file
|
|
__pycache__/tokenizer_sl.cpython-311.pyc
ADDED
Binary file (1.43 kB). View file
|
|
__pycache__/utils_sl.cpython-311.pyc
ADDED
Binary file (3.55 kB). View file
|
|
datamodule_finetune_sl.py
CHANGED
@@ -61,7 +61,7 @@ class CustomLlamaDatasetAbraham(Dataset):
|
|
61 |
return {
|
62 |
"input_ids": torch.tensor(local_encoded["input_ids"]),
|
63 |
"attention_mask": torch.tensor(local_encoded["attention_mask"]),
|
64 |
-
"labels":
|
65 |
}
|
66 |
|
67 |
class CustomFinetuneDataModule(L.LightningDataModule):
|
|
|
61 |
return {
|
62 |
"input_ids": torch.tensor(local_encoded["input_ids"]),
|
63 |
"attention_mask": torch.tensor(local_encoded["attention_mask"]),
|
64 |
+
"labels": torch.tensor(local_encoded["input_ids"]), # this one does not matter for sl
|
65 |
}
|
66 |
|
67 |
class CustomFinetuneDataModule(L.LightningDataModule):
|