wassname/peft
1
0
Fork 0
mirror of https://github.com/wassname/peft.git synced 2026-09-24 13:40:15 +08:00
Code Issues Packages Projects Releases Wiki Activity
Files
ae609d680e967b6bc9999415a88bbd30abb2ff9a
peft/examples/pet_prefix_tuning_seq2seq.ipynb
T

1964 lines
103 KiB
Plaintext
Raw Blame History

{
"cells": [
{
"cell_type": "code",
"execution_count": 1,
"id": "5f93b7d1",
"metadata": {},
"outputs": [],
"source": [
"from transformers import AutoModelForSeq2SeqLM\n",
"from pet import get_pet_config,get_pet_model, get_pet_model_state_dict\n",
"import torch\n",
"from datasets import load_dataset\n",
"import os\n",
"os.environ[\"TOKENIZERS_PARALLELISM\"] = \"false\"\n",
"os.environ[\"CUDA_VISIBLE_DEVICES\"] = \"3\"\n",
"from transformers import AutoTokenizer\n",
"from torch.utils.data import DataLoader\n",
"from transformers import default_data_collator,get_linear_schedule_with_warmup\n",
"from tqdm import tqdm\n",
"from datasets import load_dataset\n",
"\n",
"device = \"cuda\"\n",
"model_name_or_path = \"t5-large\"\n",
"tokenizer_name_or_path = \"t5-large\"\n",
"\n",
"config = {\n",
" \"pet_type\":\"PREFIX_TUNING\",\n",
" \"task_type\":\"SEQ_2_SEQ_LM\",\n",
" \"num_virtual_tokens\": 20\n",
"}\n",
"checkpoint_name = \"financial_sentiment_analysis_prefix_tuning_v1.pt\"\n",
"text_column = \"sentence\"\n",
"label_column = \"text_label\"\n",
"max_length=128\n",
"lr = 1e-2\n",
"num_epochs = 5\n",
"batch_size=8\n"
]
},
{
"cell_type": "code",
"execution_count": 2,
"id": "8d0850ac",
"metadata": {},
"outputs": [
{
"data": {
"application/vnd.jupyter.widget-view+json": {
"model_id": "3d73e8a9f76d4a2e918238f8ad0b2a30",
"version_major": 2,
"version_minor": 0
},
"text/plain": [
"Downloading: 0%| | 0.00/1.20k [00:00<?, ?B/s]"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"application/vnd.jupyter.widget-view+json": {
"model_id": "2965839a8e7f48fdafe2d7a909666b25",
"version_major": 2,
"version_minor": 0
},
"text/plain": [
"Downloading: 0%| | 0.00/2.95G [00:00<?, ?B/s]"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"trainable params: 983040 || all params: 738651136 || trainable%: 0.13308583065659835\n"
]
},
{
"data": {
"text/plain": [
"PETModelForSeq2SeqLM(\n",
" (base_model): T5ForConditionalGeneration(\n",
" (shared): Embedding(32128, 1024)\n",
" (encoder): T5Stack(\n",
" (embed_tokens): Embedding(32128, 1024)\n",
" (block): ModuleList(\n",
" (0): T5Block(\n",
" (layer): ModuleList(\n",
" (0): T5LayerSelfAttention(\n",
" (SelfAttention): T5Attention(\n",
" (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (relative_attention_bias): Embedding(32, 16)\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" (1): T5LayerFF(\n",
" (DenseReluDense): T5DenseActDense(\n",
" (wi): Linear(in_features=1024, out_features=4096, bias=False)\n",
" (wo): Linear(in_features=4096, out_features=1024, bias=False)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" (act): ReLU()\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" )\n",
" )\n",
" (1): T5Block(\n",
" (layer): ModuleList(\n",
" (0): T5LayerSelfAttention(\n",
" (SelfAttention): T5Attention(\n",
" (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" (1): T5LayerFF(\n",
" (DenseReluDense): T5DenseActDense(\n",
" (wi): Linear(in_features=1024, out_features=4096, bias=False)\n",
" (wo): Linear(in_features=4096, out_features=1024, bias=False)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" (act): ReLU()\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" )\n",
" )\n",
" (2): T5Block(\n",
" (layer): ModuleList(\n",
" (0): T5LayerSelfAttention(\n",
" (SelfAttention): T5Attention(\n",
" (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" (1): T5LayerFF(\n",
" (DenseReluDense): T5DenseActDense(\n",
" (wi): Linear(in_features=1024, out_features=4096, bias=False)\n",
" (wo): Linear(in_features=4096, out_features=1024, bias=False)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" (act): ReLU()\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" )\n",
" )\n",
" (3): T5Block(\n",
" (layer): ModuleList(\n",
" (0): T5LayerSelfAttention(\n",
" (SelfAttention): T5Attention(\n",
" (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" (1): T5LayerFF(\n",
" (DenseReluDense): T5DenseActDense(\n",
" (wi): Linear(in_features=1024, out_features=4096, bias=False)\n",
" (wo): Linear(in_features=4096, out_features=1024, bias=False)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" (act): ReLU()\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" )\n",
" )\n",
" (4): T5Block(\n",
" (layer): ModuleList(\n",
" (0): T5LayerSelfAttention(\n",
" (SelfAttention): T5Attention(\n",
" (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" (1): T5LayerFF(\n",
" (DenseReluDense): T5DenseActDense(\n",
" (wi): Linear(in_features=1024, out_features=4096, bias=False)\n",
" (wo): Linear(in_features=4096, out_features=1024, bias=False)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" (act): ReLU()\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" )\n",
" )\n",
" (5): T5Block(\n",
" (layer): ModuleList(\n",
" (0): T5LayerSelfAttention(\n",
" (SelfAttention): T5Attention(\n",
" (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" (1): T5LayerFF(\n",
" (DenseReluDense): T5DenseActDense(\n",
" (wi): Linear(in_features=1024, out_features=4096, bias=False)\n",
" (wo): Linear(in_features=4096, out_features=1024, bias=False)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" (act): ReLU()\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" )\n",
" )\n",
" (6): T5Block(\n",
" (layer): ModuleList(\n",
" (0): T5LayerSelfAttention(\n",
" (SelfAttention): T5Attention(\n",
" (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" (1): T5LayerFF(\n",
" (DenseReluDense): T5DenseActDense(\n",
" (wi): Linear(in_features=1024, out_features=4096, bias=False)\n",
" (wo): Linear(in_features=4096, out_features=1024, bias=False)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" (act): ReLU()\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" )\n",
" )\n",
" (7): T5Block(\n",
" (layer): ModuleList(\n",
" (0): T5LayerSelfAttention(\n",
" (SelfAttention): T5Attention(\n",
" (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" (1): T5LayerFF(\n",
" (DenseReluDense): T5DenseActDense(\n",
" (wi): Linear(in_features=1024, out_features=4096, bias=False)\n",
" (wo): Linear(in_features=4096, out_features=1024, bias=False)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" (act): ReLU()\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" )\n",
" )\n",
" (8): T5Block(\n",
" (layer): ModuleList(\n",
" (0): T5LayerSelfAttention(\n",
" (SelfAttention): T5Attention(\n",
" (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" (1): T5LayerFF(\n",
" (DenseReluDense): T5DenseActDense(\n",
" (wi): Linear(in_features=1024, out_features=4096, bias=False)\n",
" (wo): Linear(in_features=4096, out_features=1024, bias=False)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" (act): ReLU()\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" )\n",
" )\n",
" (9): T5Block(\n",
" (layer): ModuleList(\n",
" (0): T5LayerSelfAttention(\n",
" (SelfAttention): T5Attention(\n",
" (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" (1): T5LayerFF(\n",
" (DenseReluDense): T5DenseActDense(\n",
" (wi): Linear(in_features=1024, out_features=4096, bias=False)\n",
" (wo): Linear(in_features=4096, out_features=1024, bias=False)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" (act): ReLU()\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" )\n",
" )\n",
" (10): T5Block(\n",
" (layer): ModuleList(\n",
" (0): T5LayerSelfAttention(\n",
" (SelfAttention): T5Attention(\n",
" (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" (1): T5LayerFF(\n",
" (DenseReluDense): T5DenseActDense(\n",
" (wi): Linear(in_features=1024, out_features=4096, bias=False)\n",
" (wo): Linear(in_features=4096, out_features=1024, bias=False)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" (act): ReLU()\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" )\n",
" )\n",
" (11): T5Block(\n",
" (layer): ModuleList(\n",
" (0): T5LayerSelfAttention(\n",
" (SelfAttention): T5Attention(\n",
" (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" (1): T5LayerFF(\n",
" (DenseReluDense): T5DenseActDense(\n",
" (wi): Linear(in_features=1024, out_features=4096, bias=False)\n",
" (wo): Linear(in_features=4096, out_features=1024, bias=False)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" (act): ReLU()\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" )\n",
" )\n",
" (12): T5Block(\n",
" (layer): ModuleList(\n",
" (0): T5LayerSelfAttention(\n",
" (SelfAttention): T5Attention(\n",
" (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" (1): T5LayerFF(\n",
" (DenseReluDense): T5DenseActDense(\n",
" (wi): Linear(in_features=1024, out_features=4096, bias=False)\n",
" (wo): Linear(in_features=4096, out_features=1024, bias=False)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" (act): ReLU()\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" )\n",
" )\n",
" (13): T5Block(\n",
" (layer): ModuleList(\n",
" (0): T5LayerSelfAttention(\n",
" (SelfAttention): T5Attention(\n",
" (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" (1): T5LayerFF(\n",
" (DenseReluDense): T5DenseActDense(\n",
" (wi): Linear(in_features=1024, out_features=4096, bias=False)\n",
" (wo): Linear(in_features=4096, out_features=1024, bias=False)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" (act): ReLU()\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" )\n",
" )\n",
" (14): T5Block(\n",
" (layer): ModuleList(\n",
" (0): T5LayerSelfAttention(\n",
" (SelfAttention): T5Attention(\n",
" (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" (1): T5LayerFF(\n",
" (DenseReluDense): T5DenseActDense(\n",
" (wi): Linear(in_features=1024, out_features=4096, bias=False)\n",
" (wo): Linear(in_features=4096, out_features=1024, bias=False)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" (act): ReLU()\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" )\n",
" )\n",
" (15): T5Block(\n",
" (layer): ModuleList(\n",
" (0): T5LayerSelfAttention(\n",
" (SelfAttention): T5Attention(\n",
" (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" (1): T5LayerFF(\n",
" (DenseReluDense): T5DenseActDense(\n",
" (wi): Linear(in_features=1024, out_features=4096, bias=False)\n",
" (wo): Linear(in_features=4096, out_features=1024, bias=False)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" (act): ReLU()\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" )\n",
" )\n",
" (16): T5Block(\n",
" (layer): ModuleList(\n",
" (0): T5LayerSelfAttention(\n",
" (SelfAttention): T5Attention(\n",
" (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" (1): T5LayerFF(\n",
" (DenseReluDense): T5DenseActDense(\n",
" (wi): Linear(in_features=1024, out_features=4096, bias=False)\n",
" (wo): Linear(in_features=4096, out_features=1024, bias=False)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" (act): ReLU()\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" )\n",
" )\n",
" (17): T5Block(\n",
" (layer): ModuleList(\n",
" (0): T5LayerSelfAttention(\n",
" (SelfAttention): T5Attention(\n",
" (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" (1): T5LayerFF(\n",
" (DenseReluDense): T5DenseActDense(\n",
" (wi): Linear(in_features=1024, out_features=4096, bias=False)\n",
" (wo): Linear(in_features=4096, out_features=1024, bias=False)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" (act): ReLU()\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" )\n",
" )\n",
" (18): T5Block(\n",
" (layer): ModuleList(\n",
" (0): T5LayerSelfAttention(\n",
" (SelfAttention): T5Attention(\n",
" (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" (1): T5LayerFF(\n",
" (DenseReluDense): T5DenseActDense(\n",
" (wi): Linear(in_features=1024, out_features=4096, bias=False)\n",
" (wo): Linear(in_features=4096, out_features=1024, bias=False)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" (act): ReLU()\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" )\n",
" )\n",
" (19): T5Block(\n",
" (layer): ModuleList(\n",
" (0): T5LayerSelfAttention(\n",
" (SelfAttention): T5Attention(\n",
" (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" (1): T5LayerFF(\n",
" (DenseReluDense): T5DenseActDense(\n",
" (wi): Linear(in_features=1024, out_features=4096, bias=False)\n",
" (wo): Linear(in_features=4096, out_features=1024, bias=False)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" (act): ReLU()\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" )\n",
" )\n",
" (20): T5Block(\n",
" (layer): ModuleList(\n",
" (0): T5LayerSelfAttention(\n",
" (SelfAttention): T5Attention(\n",
" (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" (1): T5LayerFF(\n",
" (DenseReluDense): T5DenseActDense(\n",
" (wi): Linear(in_features=1024, out_features=4096, bias=False)\n",
" (wo): Linear(in_features=4096, out_features=1024, bias=False)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" (act): ReLU()\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" )\n",
" )\n",
" (21): T5Block(\n",
" (layer): ModuleList(\n",
" (0): T5LayerSelfAttention(\n",
" (SelfAttention): T5Attention(\n",
" (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" (1): T5LayerFF(\n",
" (DenseReluDense): T5DenseActDense(\n",
" (wi): Linear(in_features=1024, out_features=4096, bias=False)\n",
" (wo): Linear(in_features=4096, out_features=1024, bias=False)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" (act): ReLU()\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" )\n",
" )\n",
" (22): T5Block(\n",
" (layer): ModuleList(\n",
" (0): T5LayerSelfAttention(\n",
" (SelfAttention): T5Attention(\n",
" (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" (1): T5LayerFF(\n",
" (DenseReluDense): T5DenseActDense(\n",
" (wi): Linear(in_features=1024, out_features=4096, bias=False)\n",
" (wo): Linear(in_features=4096, out_features=1024, bias=False)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" (act): ReLU()\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" )\n",
" )\n",
" (23): T5Block(\n",
" (layer): ModuleList(\n",
" (0): T5LayerSelfAttention(\n",
" (SelfAttention): T5Attention(\n",
" (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" (1): T5LayerFF(\n",
" (DenseReluDense): T5DenseActDense(\n",
" (wi): Linear(in_features=1024, out_features=4096, bias=False)\n",
" (wo): Linear(in_features=4096, out_features=1024, bias=False)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" (act): ReLU()\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" )\n",
" )\n",
" )\n",
" (final_layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" (decoder): T5Stack(\n",
" (embed_tokens): Embedding(32128, 1024)\n",
" (block): ModuleList(\n",
" (0): T5Block(\n",
" (layer): ModuleList(\n",
" (0): T5LayerSelfAttention(\n",
" (SelfAttention): T5Attention(\n",
" (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (relative_attention_bias): Embedding(32, 16)\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" (1): T5LayerCrossAttention(\n",
" (EncDecAttention): T5Attention(\n",
" (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" (2): T5LayerFF(\n",
" (DenseReluDense): T5DenseActDense(\n",
" (wi): Linear(in_features=1024, out_features=4096, bias=False)\n",
" (wo): Linear(in_features=4096, out_features=1024, bias=False)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" (act): ReLU()\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" )\n",
" )\n",
" (1): T5Block(\n",
" (layer): ModuleList(\n",
" (0): T5LayerSelfAttention(\n",
" (SelfAttention): T5Attention(\n",
" (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" (1): T5LayerCrossAttention(\n",
" (EncDecAttention): T5Attention(\n",
" (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" (2): T5LayerFF(\n",
" (DenseReluDense): T5DenseActDense(\n",
" (wi): Linear(in_features=1024, out_features=4096, bias=False)\n",
" (wo): Linear(in_features=4096, out_features=1024, bias=False)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" (act): ReLU()\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" )\n",
" )\n",
" (2): T5Block(\n",
" (layer): ModuleList(\n",
" (0): T5LayerSelfAttention(\n",
" (SelfAttention): T5Attention(\n",
" (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" (1): T5LayerCrossAttention(\n",
" (EncDecAttention): T5Attention(\n",
" (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" (2): T5LayerFF(\n",
" (DenseReluDense): T5DenseActDense(\n",
" (wi): Linear(in_features=1024, out_features=4096, bias=False)\n",
" (wo): Linear(in_features=4096, out_features=1024, bias=False)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" (act): ReLU()\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" )\n",
" )\n",
" (3): T5Block(\n",
" (layer): ModuleList(\n",
" (0): T5LayerSelfAttention(\n",
" (SelfAttention): T5Attention(\n",
" (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" (1): T5LayerCrossAttention(\n",
" (EncDecAttention): T5Attention(\n",
" (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" (2): T5LayerFF(\n",
" (DenseReluDense): T5DenseActDense(\n",
" (wi): Linear(in_features=1024, out_features=4096, bias=False)\n",
" (wo): Linear(in_features=4096, out_features=1024, bias=False)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" (act): ReLU()\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" )\n",
" )\n",
" (4): T5Block(\n",
" (layer): ModuleList(\n",
" (0): T5LayerSelfAttention(\n",
" (SelfAttention): T5Attention(\n",
" (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" (1): T5LayerCrossAttention(\n",
" (EncDecAttention): T5Attention(\n",
" (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" (2): T5LayerFF(\n",
" (DenseReluDense): T5DenseActDense(\n",
" (wi): Linear(in_features=1024, out_features=4096, bias=False)\n",
" (wo): Linear(in_features=4096, out_features=1024, bias=False)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" (act): ReLU()\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" )\n",
" )\n",
" (5): T5Block(\n",
" (layer): ModuleList(\n",
" (0): T5LayerSelfAttention(\n",
" (SelfAttention): T5Attention(\n",
" (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" (1): T5LayerCrossAttention(\n",
" (EncDecAttention): T5Attention(\n",
" (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" (2): T5LayerFF(\n",
" (DenseReluDense): T5DenseActDense(\n",
" (wi): Linear(in_features=1024, out_features=4096, bias=False)\n",
" (wo): Linear(in_features=4096, out_features=1024, bias=False)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" (act): ReLU()\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" )\n",
" )\n",
" (6): T5Block(\n",
" (layer): ModuleList(\n",
" (0): T5LayerSelfAttention(\n",
" (SelfAttention): T5Attention(\n",
" (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" (1): T5LayerCrossAttention(\n",
" (EncDecAttention): T5Attention(\n",
" (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" (2): T5LayerFF(\n",
" (DenseReluDense): T5DenseActDense(\n",
" (wi): Linear(in_features=1024, out_features=4096, bias=False)\n",
" (wo): Linear(in_features=4096, out_features=1024, bias=False)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" (act): ReLU()\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" )\n",
" )\n",
" (7): T5Block(\n",
" (layer): ModuleList(\n",
" (0): T5LayerSelfAttention(\n",
" (SelfAttention): T5Attention(\n",
" (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" (1): T5LayerCrossAttention(\n",
" (EncDecAttention): T5Attention(\n",
" (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" (2): T5LayerFF(\n",
" (DenseReluDense): T5DenseActDense(\n",
" (wi): Linear(in_features=1024, out_features=4096, bias=False)\n",
" (wo): Linear(in_features=4096, out_features=1024, bias=False)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" (act): ReLU()\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" )\n",
" )\n",
" (8): T5Block(\n",
" (layer): ModuleList(\n",
" (0): T5LayerSelfAttention(\n",
" (SelfAttention): T5Attention(\n",
" (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" (1): T5LayerCrossAttention(\n",
" (EncDecAttention): T5Attention(\n",
" (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" (2): T5LayerFF(\n",
" (DenseReluDense): T5DenseActDense(\n",
" (wi): Linear(in_features=1024, out_features=4096, bias=False)\n",
" (wo): Linear(in_features=4096, out_features=1024, bias=False)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" (act): ReLU()\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" )\n",
" )\n",
" (9): T5Block(\n",
" (layer): ModuleList(\n",
" (0): T5LayerSelfAttention(\n",
" (SelfAttention): T5Attention(\n",
" (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" (1): T5LayerCrossAttention(\n",
" (EncDecAttention): T5Attention(\n",
" (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" (2): T5LayerFF(\n",
" (DenseReluDense): T5DenseActDense(\n",
" (wi): Linear(in_features=1024, out_features=4096, bias=False)\n",
" (wo): Linear(in_features=4096, out_features=1024, bias=False)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" (act): ReLU()\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" )\n",
" )\n",
" (10): T5Block(\n",
" (layer): ModuleList(\n",
" (0): T5LayerSelfAttention(\n",
" (SelfAttention): T5Attention(\n",
" (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" (1): T5LayerCrossAttention(\n",
" (EncDecAttention): T5Attention(\n",
" (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" (2): T5LayerFF(\n",
" (DenseReluDense): T5DenseActDense(\n",
" (wi): Linear(in_features=1024, out_features=4096, bias=False)\n",
" (wo): Linear(in_features=4096, out_features=1024, bias=False)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" (act): ReLU()\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" )\n",
" )\n",
" (11): T5Block(\n",
" (layer): ModuleList(\n",
" (0): T5LayerSelfAttention(\n",
" (SelfAttention): T5Attention(\n",
" (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" (1): T5LayerCrossAttention(\n",
" (EncDecAttention): T5Attention(\n",
" (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" (2): T5LayerFF(\n",
" (DenseReluDense): T5DenseActDense(\n",
" (wi): Linear(in_features=1024, out_features=4096, bias=False)\n",
" (wo): Linear(in_features=4096, out_features=1024, bias=False)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" (act): ReLU()\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" )\n",
" )\n",
" (12): T5Block(\n",
" (layer): ModuleList(\n",
" (0): T5LayerSelfAttention(\n",
" (SelfAttention): T5Attention(\n",
" (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" (1): T5LayerCrossAttention(\n",
" (EncDecAttention): T5Attention(\n",
" (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" (2): T5LayerFF(\n",
" (DenseReluDense): T5DenseActDense(\n",
" (wi): Linear(in_features=1024, out_features=4096, bias=False)\n",
" (wo): Linear(in_features=4096, out_features=1024, bias=False)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" (act): ReLU()\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" )\n",
" )\n",
" (13): T5Block(\n",
" (layer): ModuleList(\n",
" (0): T5LayerSelfAttention(\n",
" (SelfAttention): T5Attention(\n",
" (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" (1): T5LayerCrossAttention(\n",
" (EncDecAttention): T5Attention(\n",
" (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" (2): T5LayerFF(\n",
" (DenseReluDense): T5DenseActDense(\n",
" (wi): Linear(in_features=1024, out_features=4096, bias=False)\n",
" (wo): Linear(in_features=4096, out_features=1024, bias=False)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" (act): ReLU()\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" )\n",
" )\n",
" (14): T5Block(\n",
" (layer): ModuleList(\n",
" (0): T5LayerSelfAttention(\n",
" (SelfAttention): T5Attention(\n",
" (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" (1): T5LayerCrossAttention(\n",
" (EncDecAttention): T5Attention(\n",
" (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" (2): T5LayerFF(\n",
" (DenseReluDense): T5DenseActDense(\n",
" (wi): Linear(in_features=1024, out_features=4096, bias=False)\n",
" (wo): Linear(in_features=4096, out_features=1024, bias=False)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" (act): ReLU()\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" )\n",
" )\n",
" (15): T5Block(\n",
" (layer): ModuleList(\n",
" (0): T5LayerSelfAttention(\n",
" (SelfAttention): T5Attention(\n",
" (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" (1): T5LayerCrossAttention(\n",
" (EncDecAttention): T5Attention(\n",
" (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" (2): T5LayerFF(\n",
" (DenseReluDense): T5DenseActDense(\n",
" (wi): Linear(in_features=1024, out_features=4096, bias=False)\n",
" (wo): Linear(in_features=4096, out_features=1024, bias=False)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" (act): ReLU()\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" )\n",
" )\n",
" (16): T5Block(\n",
" (layer): ModuleList(\n",
" (0): T5LayerSelfAttention(\n",
" (SelfAttention): T5Attention(\n",
" (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" (1): T5LayerCrossAttention(\n",
" (EncDecAttention): T5Attention(\n",
" (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" (2): T5LayerFF(\n",
" (DenseReluDense): T5DenseActDense(\n",
" (wi): Linear(in_features=1024, out_features=4096, bias=False)\n",
" (wo): Linear(in_features=4096, out_features=1024, bias=False)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" (act): ReLU()\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" )\n",
" )\n",
" (17): T5Block(\n",
" (layer): ModuleList(\n",
" (0): T5LayerSelfAttention(\n",
" (SelfAttention): T5Attention(\n",
" (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" (1): T5LayerCrossAttention(\n",
" (EncDecAttention): T5Attention(\n",
" (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" (2): T5LayerFF(\n",
" (DenseReluDense): T5DenseActDense(\n",
" (wi): Linear(in_features=1024, out_features=4096, bias=False)\n",
" (wo): Linear(in_features=4096, out_features=1024, bias=False)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" (act): ReLU()\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" )\n",
" )\n",
" (18): T5Block(\n",
" (layer): ModuleList(\n",
" (0): T5LayerSelfAttention(\n",
" (SelfAttention): T5Attention(\n",
" (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" (1): T5LayerCrossAttention(\n",
" (EncDecAttention): T5Attention(\n",
" (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" (2): T5LayerFF(\n",
" (DenseReluDense): T5DenseActDense(\n",
" (wi): Linear(in_features=1024, out_features=4096, bias=False)\n",
" (wo): Linear(in_features=4096, out_features=1024, bias=False)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" (act): ReLU()\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" )\n",
" )\n",
" (19): T5Block(\n",
" (layer): ModuleList(\n",
" (0): T5LayerSelfAttention(\n",
" (SelfAttention): T5Attention(\n",
" (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" (1): T5LayerCrossAttention(\n",
" (EncDecAttention): T5Attention(\n",
" (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" (2): T5LayerFF(\n",
" (DenseReluDense): T5DenseActDense(\n",
" (wi): Linear(in_features=1024, out_features=4096, bias=False)\n",
" (wo): Linear(in_features=4096, out_features=1024, bias=False)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" (act): ReLU()\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" )\n",
" )\n",
" (20): T5Block(\n",
" (layer): ModuleList(\n",
" (0): T5LayerSelfAttention(\n",
" (SelfAttention): T5Attention(\n",
" (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" (1): T5LayerCrossAttention(\n",
" (EncDecAttention): T5Attention(\n",
" (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" (2): T5LayerFF(\n",
" (DenseReluDense): T5DenseActDense(\n",
" (wi): Linear(in_features=1024, out_features=4096, bias=False)\n",
" (wo): Linear(in_features=4096, out_features=1024, bias=False)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" (act): ReLU()\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" )\n",
" )\n",
" (21): T5Block(\n",
" (layer): ModuleList(\n",
" (0): T5LayerSelfAttention(\n",
" (SelfAttention): T5Attention(\n",
" (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" (1): T5LayerCrossAttention(\n",
" (EncDecAttention): T5Attention(\n",
" (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" (2): T5LayerFF(\n",
" (DenseReluDense): T5DenseActDense(\n",
" (wi): Linear(in_features=1024, out_features=4096, bias=False)\n",
" (wo): Linear(in_features=4096, out_features=1024, bias=False)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" (act): ReLU()\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" )\n",
" )\n",
" (22): T5Block(\n",
" (layer): ModuleList(\n",
" (0): T5LayerSelfAttention(\n",
" (SelfAttention): T5Attention(\n",
" (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" (1): T5LayerCrossAttention(\n",
" (EncDecAttention): T5Attention(\n",
" (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" (2): T5LayerFF(\n",
" (DenseReluDense): T5DenseActDense(\n",
" (wi): Linear(in_features=1024, out_features=4096, bias=False)\n",
" (wo): Linear(in_features=4096, out_features=1024, bias=False)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" (act): ReLU()\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" )\n",
" )\n",
" (23): T5Block(\n",
" (layer): ModuleList(\n",
" (0): T5LayerSelfAttention(\n",
" (SelfAttention): T5Attention(\n",
" (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" (1): T5LayerCrossAttention(\n",
" (EncDecAttention): T5Attention(\n",
" (q): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (k): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (v): Linear(in_features=1024, out_features=1024, bias=False)\n",
" (o): Linear(in_features=1024, out_features=1024, bias=False)\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" (2): T5LayerFF(\n",
" (DenseReluDense): T5DenseActDense(\n",
" (wi): Linear(in_features=1024, out_features=4096, bias=False)\n",
" (wo): Linear(in_features=4096, out_features=1024, bias=False)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" (act): ReLU()\n",
" )\n",
" (layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" )\n",
" )\n",
" )\n",
" (final_layer_norm): FusedRMSNorm(torch.Size([1024]), eps=1e-06, elementwise_affine=True)\n",
" (dropout): Dropout(p=0.1, inplace=False)\n",
" )\n",
" (lm_head): Linear(in_features=1024, out_features=32128, bias=False)\n",
" )\n",
" (word_embeddings): Embedding(32128, 1024)\n",
" (prompt_encoder): PrefixEncoder(\n",
" (embedding): Embedding(20, 49152)\n",
" )\n",
")"
]
},
"execution_count": 2,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"# creating model\n",
"pet_config = get_pet_config(config)\n",
"\n",
"model = AutoModelForSeq2SeqLM.from_pretrained(model_name_or_path)\n",
"model = get_pet_model(model, pet_config)\n",
"model.print_trainable_parameters()\n",
"model"
]
},
{
"cell_type": "code",
"execution_count": 3,
"id": "4ee2babf",
"metadata": {},
"outputs": [
{
"name": "stderr",
"output_type": "stream",
"text": [
"/home/sourab/miniconda3/envs/ml/lib/python3.10/site-packages/huggingface_hub/utils/_deprecation.py:97: FutureWarning: Deprecated argument(s) used in 'dataset_info': token. Will not be supported from version '0.12'.\n",
" warnings.warn(message, FutureWarning)\n",
"Found cached dataset financial_phrasebank (/home/sourab/.cache/huggingface/datasets/financial_phrasebank/sentences_allagree/1.0.0/550bde12e6c30e2674da973a55f57edde5181d53f5a5a34c1531c53f93b7e141)\n"
]
},
{
"data": {
"application/vnd.jupyter.widget-view+json": {
"model_id": "e3f8b8faca0a4112b2c3499faee9544b",
"version_major": 2,
"version_minor": 0
},
"text/plain": [
" 0%| | 0/1 [00:00<?, ?it/s]"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"application/vnd.jupyter.widget-view+json": {
"model_id": "935c8aebde284a5784348588e0bb013a",
"version_major": 2,
"version_minor": 0
},
"text/plain": [
" 0%| | 0/3 [00:00<?, ?ba/s]"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"application/vnd.jupyter.widget-view+json": {
"model_id": "e3487cd55f6847588492bf7fa51348ca",
"version_major": 2,
"version_minor": 0
},
"text/plain": [
" 0%| | 0/1 [00:00<?, ?ba/s]"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"text/plain": [
"{'sentence': 'ADPnews - Feb 5 , 2010 - Finnish real estate investor Sponda Oyj HEL : SDA1V said today that it slipped to a net loss of EUR 81.5 million USD 11.8 m in 2009 from a profit of EUR 29.3 million in 2008 .',\n",
" 'label': 0,\n",
" 'text_label': 'negative'}"
]
},
"execution_count": 3,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"# loading dataset\n",
"dataset = load_dataset(\"financial_phrasebank\", 'sentences_allagree')\n",
"dataset = dataset[\"train\"].train_test_split(test_size=0.1)\n",
"dataset[\"validation\"] = dataset[\"test\"]\n",
"del(dataset[\"test\"])\n",
"\n",
"classes = dataset[\"train\"].features[\"label\"].names\n",
"dataset = dataset.map(\n",
" lambda x: {\"text_label\": [classes[label] for label in x[\"label\"]]},\n",
" batched=True,\n",
" num_proc=1,\n",
" \n",
")\n",
"\n",
"dataset[\"train\"][0]"
]
},
{
"cell_type": "code",
"execution_count": 4,
"id": "adf9608c",
"metadata": {},
"outputs": [
{
"data": {
"application/vnd.jupyter.widget-view+json": {
"model_id": "2ce088f4437d4e2c80c267332a5b84e5",
"version_major": 2,
"version_minor": 0
},
"text/plain": [
"Downloading: 0%| | 0.00/792k [00:00<?, ?B/s]"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"application/vnd.jupyter.widget-view+json": {
"model_id": "4e5f69b61f194220b39336e48edd2f9e",
"version_major": 2,
"version_minor": 0
},
"text/plain": [
"Downloading: 0%| | 0.00/1.39M [00:00<?, ?B/s]"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"/home/sourab/transformers/src/transformers/models/t5/tokenization_t5_fast.py:156: FutureWarning: This tokenizer was incorrectly instantiated with a model max length of 512 which will be corrected in Transformers v5.\n",
"For now, this behavior is kept to avoid breaking backwards compatibility when padding/encoding with `truncation is True`.\n",
"- Be aware that you SHOULD NOT rely on t5-large automatically truncating your input to 512 when padding/encoding.\n",
"- If you want to encode/pad to sequences longer than 512 you can either instantiate this tokenizer with `model_max_length` or pass `max_length` when encoding/padding.\n",
"- To avoid this warning, please instantiate this tokenizer with `model_max_length` set to your preferred value.\n",
" warnings.warn(\n"
]
},
{
"data": {
"application/vnd.jupyter.widget-view+json": {
"model_id": "230c5631891e4ea8ac7a1b39f315a4f0",
"version_major": 2,
"version_minor": 0
},
"text/plain": [
"Running tokenizer on dataset: 0%| | 0/3 [00:00<?, ?ba/s]"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"application/vnd.jupyter.widget-view+json": {
"model_id": "b581e5677d2a45459ceb725534ed0891",
"version_major": 2,
"version_minor": 0
},
"text/plain": [
"Running tokenizer on dataset: 0%| | 0/1 [00:00<?, ?ba/s]"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"# data preprocessing\n",
"tokenizer = AutoTokenizer.from_pretrained(model_name_or_path)\n",
"def preprocess_function(examples):\n",
" inputs = examples[text_column]\n",
" targets = examples[label_column]\n",
" model_inputs = tokenizer(inputs, max_length=max_length, padding=\"max_length\", truncation=True, return_tensors=\"pt\")\n",
" labels = tokenizer(targets, max_length=2, padding=\"max_length\", truncation=True, return_tensors=\"pt\")\n",
" labels = labels[\"input_ids\"]\n",
" labels[labels==tokenizer.pad_token_id] = -100\n",
" model_inputs[\"labels\"] = labels\n",
" return model_inputs\n",
"\n",
"processed_datasets = dataset.map(\n",
" preprocess_function,\n",
" batched=True,\n",
" num_proc=1,\n",
" remove_columns=dataset[\"train\"].column_names,\n",
" load_from_cache_file=False,\n",
" desc=\"Running tokenizer on dataset\",\n",
" )\n",
"\n",
"train_dataset = processed_datasets[\"train\"]\n",
"eval_dataset = processed_datasets[\"validation\"]\n",
"\n",
"train_dataloader = DataLoader(\n",
" train_dataset, shuffle=True, collate_fn=default_data_collator, batch_size=batch_size, pin_memory=True\n",
" )\n",
"eval_dataloader = DataLoader(eval_dataset, collate_fn=default_data_collator, batch_size=batch_size, pin_memory=True)\n",
"\n",
"\n",
"\n",
" "
]
},
{
"cell_type": "code",
"execution_count": 5,
"id": "f733a3c6",
"metadata": {},
"outputs": [],
"source": [
"# optimizer and lr scheduler\n",
"optimizer = torch.optim.AdamW(model.parameters(), lr=lr)\n",
"lr_scheduler = get_linear_schedule_with_warmup(\n",
" optimizer=optimizer,\n",
" num_warmup_steps=0,\n",
" num_training_steps=(len(train_dataloader) * num_epochs),\n",
")\n"
]
},
{
"cell_type": "code",
"execution_count": 6,
"id": "6b3a4090",
"metadata": {},
"outputs": [
{
"name": "stderr",
"output_type": "stream",
"text": [
"100%|█████████████████████████████████████████████████████████████| 255/255 [00:20<00:00, 12.27it/s]\n",
"100%|███████████████████████████████████████████████████████████████| 29/29 [00:01<00:00, 17.32it/s]\n"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"epoch=0: train_ppl=tensor(2697769., device='cuda:0') train_epoch_loss=tensor(14.8079, device='cuda:0') eval_ppl=tensor(1.0089, device='cuda:0') eval_epoch_loss=tensor(0.0089, device='cuda:0')\n"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"100%|█████████████████████████████████████████████████████████████| 255/255 [00:19<00:00, 12.75it/s]\n",
"100%|███████████████████████████████████████████████████████████████| 29/29 [00:01<00:00, 17.33it/s]\n"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"epoch=1: train_ppl=tensor(2.9475, device='cuda:0') train_epoch_loss=tensor(1.0809, device='cuda:0') eval_ppl=tensor(1.0072, device='cuda:0') eval_epoch_loss=tensor(0.0072, device='cuda:0')\n"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"100%|█████████████████████████████████████████████████████████████| 255/255 [00:20<00:00, 12.71it/s]\n",
"100%|███████████████████████████████████████████████████████████████| 29/29 [00:01<00:00, 17.31it/s]\n"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"epoch=2: train_ppl=tensor(2.0588, device='cuda:0') train_epoch_loss=tensor(0.7221, device='cuda:0') eval_ppl=tensor(1.0055, device='cuda:0') eval_epoch_loss=tensor(0.0054, device='cuda:0')\n"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"100%|█████████████████████████████████████████████████████████████| 255/255 [00:20<00:00, 12.70it/s]\n",
"100%|███████████████████████████████████████████████████████████████| 29/29 [00:01<00:00, 17.32it/s]\n"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"epoch=3: train_ppl=tensor(1.7939, device='cuda:0') train_epoch_loss=tensor(0.5844, device='cuda:0') eval_ppl=tensor(1.0063, device='cuda:0') eval_epoch_loss=tensor(0.0063, device='cuda:0')\n"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"100%|█████████████████████████████████████████████████████████████| 255/255 [00:19<00:00, 13.01it/s]\n",
"100%|███████████████████████████████████████████████████████████████| 29/29 [00:01<00:00, 17.33it/s]"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"epoch=4: train_ppl=tensor(1.7740, device='cuda:0') train_epoch_loss=tensor(0.5732, device='cuda:0') eval_ppl=tensor(1.0062, device='cuda:0') eval_epoch_loss=tensor(0.0061, device='cuda:0')\n"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"\n"
]
}
],
"source": [
"# training and evaluation\n",
"model = model.to(device)\n",
"\n",
"for epoch in range(num_epochs):\n",
" model.train()\n",
" total_loss = 0\n",
" for step, batch in enumerate(tqdm(train_dataloader)):\n",
" batch = {k: v.to(device) for k, v in batch.items()}\n",
" outputs = model(**batch)\n",
" loss = outputs.loss\n",
" total_loss += loss.detach().float()\n",
" loss.backward()\n",
" optimizer.step()\n",
" lr_scheduler.step()\n",
" optimizer.zero_grad()\n",
"\n",
" model.eval()\n",
" eval_loss = 0\n",
" eval_preds = []\n",
" for step, batch in enumerate(tqdm(eval_dataloader)):\n",
" batch = {k: v.to(device) for k, v in batch.items()}\n",
" with torch.no_grad():\n",
" outputs = model(**batch)\n",
" loss = outputs.loss\n",
" eval_loss += loss.detach().float()\n",
" eval_preds.extend(tokenizer.batch_decode(torch.argmax(outputs.logits, -1).detach().cpu().numpy(), skip_special_tokens=True))\n",
"\n",
" eval_epoch_loss = eval_loss/len(train_dataloader)\n",
" eval_ppl = torch.exp(eval_epoch_loss)\n",
" train_epoch_loss = total_loss/len(eval_dataloader)\n",
" train_ppl = torch.exp(train_epoch_loss)\n",
" print(f\"{epoch=}: {train_ppl=} {train_epoch_loss=} {eval_ppl=} {eval_epoch_loss=}\")\n"
]
},
{
"cell_type": "code",
"execution_count": 7,
"id": "6cafa67b",
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"accuracy=96.47577092511013 % on the evaluation dataset\n",
"eval_preds[:10]=['neutral', 'neutral', 'neutral', 'negative', 'neutral', 'neutral', 'neutral', 'neutral', 'positive', 'positive']\n",
"dataset['validation']['text_label'][:10]=['neutral', 'neutral', 'neutral', 'negative', 'neutral', 'neutral', 'neutral', 'neutral', 'positive', 'positive']\n"
]
}
],
"source": [
"# print accuracy\n",
"correct =0\n",
"total = 0\n",
"for pred,true in zip(eval_preds, dataset[\"validation\"][\"text_label\"]):\n",
" if pred.strip()==true.strip():\n",
" correct+=1\n",
" total+=1 \n",
"accuracy = correct/total*100\n",
"print(f\"{accuracy=} % on the evaluation dataset\")\n",
"print(f\"{eval_preds[:10]=}\")\n",
"print(f\"{dataset['validation']['text_label'][:10]=}\")"
]
},
{
"cell_type": "code",
"execution_count": 8,
"id": "a8de6005",
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"{'prompt_embeddings': tensor([[-0.3165, -0.8389, 0.3262, ..., -1.5049, -1.6963, 0.3444],\n",
" [-1.8359, 1.1936, 1.0483, ..., 0.6197, -0.4452, 0.5844],\n",
" [-0.6027, 0.3246, -1.5601, ..., -0.3645, 0.2329, 0.3402],\n",
" ...,\n",
" [-1.9525, -0.5035, 0.8474, ..., 0.4793, -0.0789, -0.9305],\n",
" [-1.9741, 0.5242, -2.0594, ..., -0.7970, -0.4889, 2.7323],\n",
" [ 0.9355, -0.2714, 0.4610, ..., 0.2692, -1.5801, -1.6405]])}\n"
]
}
],
"source": [
"# saving model\n",
"state_dict = get_pet_model_state_dict(model)\n",
"torch.save(state_dict, checkpoint_name)\n",
"print(state_dict)"
]
},
{
"cell_type": "code",
"execution_count": 9,
"id": "bd20cd4c",
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"3,8M\tfinancial_sentiment_analysis_prefix_tuning_v1.pt\r\n"
]
}
],
"source": [
"!du -h $checkpoint_name"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "76c2fc29",
"metadata": {},
"outputs": [],
"source": []
}
],
"metadata": {
"kernelspec": {
"display_name": "Python 3 (ipykernel)",
"language": "python",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.10.4"
}
},
"nbformat": 4,
"nbformat_minor": 5
}
Reference in New Issue View Git Blame Copy Permalink