mirror of
https://github.com/ltcptgeneral/cse151b-final-project.git
synced 2026-10-11 23:28:31 +00:00
removed test stuff
This commit is contained in:
Ethan Shapiro committed
1 parent
7ad5b97463
commit
8d3ce990e3
4 files changed
+116
-227
No files matched your search
+2
-1
@@ -1,2 +1,3 @@
|
||||
**/data/*
|
||||
/env
|
||||
/env
|
||||
**/*.zip
|
||||
@@ -0,0 +1,114 @@
|
||||
{
|
||||
"cells": [
|
||||
{
|
||||
"cell_type": "code",
|
||||
"execution_count": null,
|
||||
"metadata": {},
|
||||
"outputs": [],
|
||||
"source": [
|
||||
"import gym\n",
|
||||
"import gym_wordle\n",
|
||||
"from stable_baselines3 import DQN\n",
|
||||
"import numpy as np\n",
|
||||
"import tqdm"
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"execution_count": null,
|
||||
"metadata": {},
|
||||
"outputs": [],
|
||||
"source": [
|
||||
"env = gym.make(\"Wordle-v0\")\n",
|
||||
"\n",
|
||||
"print(env)"
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"execution_count": 35,
|
||||
"metadata": {},
|
||||
"outputs": [],
|
||||
"source": [
|
||||
"total_timesteps = 100000\n",
|
||||
"model = DQN(\"MlpPolicy\", env, verbose=0)\n",
|
||||
"model.learn(total_timesteps=total_timesteps, progress_bar=True)"
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"execution_count": null,
|
||||
"metadata": {},
|
||||
"outputs": [],
|
||||
"source": [
|
||||
"def test(model):\n",
|
||||
"\n",
|
||||
" end_rewards = []\n",
|
||||
"\n",
|
||||
" for i in range(1000):\n",
|
||||
" \n",
|
||||
" state = env.reset()\n",
|
||||
"\n",
|
||||
" done = False\n",
|
||||
"\n",
|
||||
" while not done:\n",
|
||||
"\n",
|
||||
" action, _states = model.predict(state, deterministic=True)\n",
|
||||
"\n",
|
||||
" state, reward, done, info = env.step(action)\n",
|
||||
" \n",
|
||||
" end_rewards.append(reward == 0)\n",
|
||||
" \n",
|
||||
" return np.sum(end_rewards) / len(end_rewards)"
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"execution_count": null,
|
||||
"metadata": {},
|
||||
"outputs": [],
|
||||
"source": [
|
||||
"model.save(\"dqn_wordle\")"
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"execution_count": null,
|
||||
"metadata": {},
|
||||
"outputs": [],
|
||||
"source": [
|
||||
"model = DQN.load(\"dqn_wordle\")"
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"execution_count": null,
|
||||
"metadata": {},
|
||||
"outputs": [],
|
||||
"source": [
|
||||
"print(test(model))"
|
||||
]
|
||||
}
|
||||
],
|
||||
"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": "3.8.10"
|
||||
}
|
||||
},
|
||||
"nbformat": 4,
|
||||
"nbformat_minor": 2
|
||||
}
|
||||
-165
@@ -1,165 +0,0 @@
|
||||
{
|
||||
"cells": [
|
||||
{
|
||||
"cell_type": "code",
|
||||
"execution_count": 1,
|
||||
"metadata": {},
|
||||
"outputs": [
|
||||
{
|
||||
"name": "stderr",
|
||||
"output_type": "stream",
|
||||
"text": [
|
||||
"C:\\Users\\art\\AppData\\Local\\Packages\\PythonSoftwareFoundation.Python.3.11_qbz5n2kfra8p0\\LocalCache\\local-packages\\Python311\\site-packages\\tqdm\\auto.py:21: TqdmWarning: IProgress not found. Please update jupyter and ipywidgets. See https://ipywidgets.readthedocs.io/en/stable/user_install.html\n",
|
||||
" from .autonotebook import tqdm as notebook_tqdm\n"
|
||||
]
|
||||
}
|
||||
],
|
||||
"source": [
|
||||
"from torch.utils.data import Dataset\n",
|
||||
"from transformers import BertGenerationEncoder, BertGenerationDecoder, EncoderDecoderModel, BertTokenizer\n",
|
||||
"from tqdm import tqdm as progress_bar\n",
|
||||
"import torch\n",
|
||||
"import matplotlib"
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"execution_count": 2,
|
||||
"metadata": {},
|
||||
"outputs": [
|
||||
{
|
||||
"name": "stdout",
|
||||
"output_type": "stream",
|
||||
"text": [
|
||||
"cuda\n"
|
||||
]
|
||||
}
|
||||
],
|
||||
"source": [
|
||||
"device = torch.device(\"cuda\" if torch.cuda.is_available() else \"cpu\")\n",
|
||||
"print(device)"
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"execution_count": 3,
|
||||
"metadata": {},
|
||||
"outputs": [
|
||||
{
|
||||
"name": "stderr",
|
||||
"output_type": "stream",
|
||||
"text": [
|
||||
"You are using a model of type bert to instantiate a model of type bert-generation. This is not supported for all configurations of models and can yield errors.\n",
|
||||
"You are using a model of type bert to instantiate a model of type bert-generation. This is not supported for all configurations of models and can yield errors.\n",
|
||||
"Some weights of BertGenerationDecoder were not initialized from the model checkpoint at google-bert/bert-large-uncased and are newly initialized: ['bert.encoder.layer.0.crossattention.output.LayerNorm.bias', 'bert.encoder.layer.0.crossattention.output.LayerNorm.weight', 'bert.encoder.layer.0.crossattention.output.dense.bias', 'bert.encoder.layer.0.crossattention.output.dense.weight', 'bert.encoder.layer.0.crossattention.self.key.bias', 'bert.encoder.layer.0.crossattention.self.key.weight', 'bert.encoder.layer.0.crossattention.self.query.bias', 'bert.encoder.layer.0.crossattention.self.query.weight', 'bert.encoder.layer.0.crossattention.self.value.bias', 'bert.encoder.layer.0.crossattention.self.value.weight', 'bert.encoder.layer.1.crossattention.output.LayerNorm.bias', 'bert.encoder.layer.1.crossattention.output.LayerNorm.weight', 'bert.encoder.layer.1.crossattention.output.dense.bias', 'bert.encoder.layer.1.crossattention.output.dense.weight', 'bert.encoder.layer.1.crossattention.self.key.bias', 'bert.encoder.layer.1.crossattention.self.key.weight', 'bert.encoder.layer.1.crossattention.self.query.bias', 'bert.encoder.layer.1.crossattention.self.query.weight', 'bert.encoder.layer.1.crossattention.self.value.bias', 'bert.encoder.layer.1.crossattention.self.value.weight', 'bert.encoder.layer.10.crossattention.output.LayerNorm.bias', 'bert.encoder.layer.10.crossattention.output.LayerNorm.weight', 'bert.encoder.layer.10.crossattention.output.dense.bias', 'bert.encoder.layer.10.crossattention.output.dense.weight', 'bert.encoder.layer.10.crossattention.self.key.bias', 'bert.encoder.layer.10.crossattention.self.key.weight', 'bert.encoder.layer.10.crossattention.self.query.bias', 'bert.encoder.layer.10.crossattention.self.query.weight', 'bert.encoder.layer.10.crossattention.self.value.bias', 'bert.encoder.layer.10.crossattention.self.value.weight', 'bert.encoder.layer.11.crossattention.output.LayerNorm.bias', 'bert.encoder.layer.11.crossattention.output.LayerNorm.weight', 'bert.encoder.layer.11.crossattention.output.dense.bias', 'bert.encoder.layer.11.crossattention.output.dense.weight', 'bert.encoder.layer.11.crossattention.self.key.bias', 'bert.encoder.layer.11.crossattention.self.key.weight', 'bert.encoder.layer.11.crossattention.self.query.bias', 'bert.encoder.layer.11.crossattention.self.query.weight', 'bert.encoder.layer.11.crossattention.self.value.bias', 'bert.encoder.layer.11.crossattention.self.value.weight', 'bert.encoder.layer.12.crossattention.output.LayerNorm.bias', 'bert.encoder.layer.12.crossattention.output.LayerNorm.weight', 'bert.encoder.layer.12.crossattention.output.dense.bias', 'bert.encoder.layer.12.crossattention.output.dense.weight', 'bert.encoder.layer.12.crossattention.self.key.bias', 'bert.encoder.layer.12.crossattention.self.key.weight', 'bert.encoder.layer.12.crossattention.self.query.bias', 'bert.encoder.layer.12.crossattention.self.query.weight', 'bert.encoder.layer.12.crossattention.self.value.bias', 'bert.encoder.layer.12.crossattention.self.value.weight', 'bert.encoder.layer.13.crossattention.output.LayerNorm.bias', 'bert.encoder.layer.13.crossattention.output.LayerNorm.weight', 'bert.encoder.layer.13.crossattention.output.dense.bias', 'bert.encoder.layer.13.crossattention.output.dense.weight', 'bert.encoder.layer.13.crossattention.self.key.bias', 'bert.encoder.layer.13.crossattention.self.key.weight', 'bert.encoder.layer.13.crossattention.self.query.bias', 'bert.encoder.layer.13.crossattention.self.query.weight', 'bert.encoder.layer.13.crossattention.self.value.bias', 'bert.encoder.layer.13.crossattention.self.value.weight', 'bert.encoder.layer.14.crossattention.output.LayerNorm.bias', 'bert.encoder.layer.14.crossattention.output.LayerNorm.weight', 'bert.encoder.layer.14.crossattention.output.dense.bias', 'bert.encoder.layer.14.crossattention.output.dense.weight', 'bert.encoder.layer.14.crossattention.self.key.bias', 'bert.encoder.layer.14.crossattention.self.key.weight', 'bert.encoder.layer.14.crossattention.self.query.bias', 'bert.encoder.layer.14.crossattention.self.query.weight', 'bert.encoder.layer.14.crossattention.self.value.bias', 'bert.encoder.layer.14.crossattention.self.value.weight', 'bert.encoder.layer.15.crossattention.output.LayerNorm.bias', 'bert.encoder.layer.15.crossattention.output.LayerNorm.weight', 'bert.encoder.layer.15.crossattention.output.dense.bias', 'bert.encoder.layer.15.crossattention.output.dense.weight', 'bert.encoder.layer.15.crossattention.self.key.bias', 'bert.encoder.layer.15.crossattention.self.key.weight', 'bert.encoder.layer.15.crossattention.self.query.bias', 'bert.encoder.layer.15.crossattention.self.query.weight', 'bert.encoder.layer.15.crossattention.self.value.bias', 'bert.encoder.layer.15.crossattention.self.value.weight', 'bert.encoder.layer.16.crossattention.output.LayerNorm.bias', 'bert.encoder.layer.16.crossattention.output.LayerNorm.weight', 'bert.encoder.layer.16.crossattention.output.dense.bias', 'bert.encoder.layer.16.crLine truncated
|
||||
"You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.\n"
|
||||
]
|
||||
}
|
||||
],
|
||||
"source": [
|
||||
"encoder = BertGenerationEncoder.from_pretrained(\"google-bert/bert-large-uncased\", bos_token_id=101, eos_token_id=102)\n",
|
||||
"# add cross attention layers and use BERT's cls token as BOS token and sep token as EOS token\n",
|
||||
"decoder = BertGenerationDecoder.from_pretrained(\"google-bert/bert-large-uncased\", add_cross_attention=True, is_decoder=True, bos_token_id=101, eos_token_id=102)\n",
|
||||
"model = EncoderDecoderModel(encoder=encoder, decoder=decoder)\n",
|
||||
"\n",
|
||||
"# create tokenizer...\n",
|
||||
"tokenizer = BertTokenizer.from_pretrained(\"google-bert/bert-large-uncased\")"
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"execution_count": 5,
|
||||
"metadata": {},
|
||||
"outputs": [],
|
||||
"source": [
|
||||
"import json\n",
|
||||
"\n",
|
||||
"class CodeDataset(Dataset):\n",
|
||||
" def __init__(self):\n",
|
||||
" with open(\"data/conala-train.json\") as f:\n",
|
||||
" self.data = json.load(f)\n",
|
||||
"\n",
|
||||
" def __len__(self):\n",
|
||||
" return len(self.data)\n",
|
||||
"\n",
|
||||
" def __getitem__(self, idx):\n",
|
||||
" return self.data[idx][\"rewritten_intent\"], self.data[idx][\"snippet\"]"
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"execution_count": 6,
|
||||
"metadata": {},
|
||||
"outputs": [],
|
||||
"source": [
|
||||
"optimizer = torch.optim.AdamW(params=model.parameters(), lr=1e-3)\n",
|
||||
"dataloader = CodeDataset()\n",
|
||||
"model = model.to(device)"
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"execution_count": 7,
|
||||
"metadata": {},
|
||||
"outputs": [
|
||||
{
|
||||
"name": "stderr",
|
||||
"output_type": "stream",
|
||||
"text": [
|
||||
" 0%| | 0/2379 [00:00<?, ?it/s]C:\\Users\\art\\AppData\\Local\\Packages\\PythonSoftwareFoundation.Python.3.11_qbz5n2kfra8p0\\LocalCache\\local-packages\\Python311\\site-packages\\transformers\\models\\encoder_decoder\\modeling_encoder_decoder.py:636: FutureWarning: Version v4.12.0 introduces a better way to train encoder-decoder models by computing the loss inside the encoder-decoder framework rather than in the decoder itself. You may observe training discrepancies if fine-tuning a model trained with versions anterior to 4.12.0. The decoder_input_ids are now created based on the labels, no need to pass them yourself anymore.\n",
|
||||
" warnings.warn(DEPRECATION_WARNING, FutureWarning)\n",
|
||||
" 0%| | 3/2379 [01:18<16:16:50, 24.67s/it]"
|
||||
]
|
||||
}
|
||||
],
|
||||
"source": [
|
||||
"losses = []\n",
|
||||
"epochs = 10\n",
|
||||
"for i in range(epochs):\n",
|
||||
"\n",
|
||||
" epoch_loss = 0\n",
|
||||
"\n",
|
||||
" for idx, (question, answer) in progress_bar(enumerate(dataloader), total=len(dataloader)):\n",
|
||||
"\n",
|
||||
" print(questions)\n",
|
||||
" input_ids = tokenizer(question, add_special_tokens=False, return_tensors=\"pt\").input_ids.to(device)\n",
|
||||
" label_ids = tokenizer(answer, return_tensors=\"pt\").input_ids.to(device)\n",
|
||||
"\n",
|
||||
" loss = model(input_ids=input_ids, decoder_input_ids=label_ids, labels=label_ids).loss\n",
|
||||
"\n",
|
||||
" optimizer.zero_grad()\n",
|
||||
" loss.backward()\n",
|
||||
" optimizer.step()\n",
|
||||
"\n",
|
||||
" epoch_loss += loss.item()\n",
|
||||
"\n",
|
||||
" losses.append(epoch_loss)\n",
|
||||
"\n",
|
||||
"plt.plot(losses, color=\"green\", label=\"Training Loss\")\n",
|
||||
"plt.legend(loc = 'upper left')\n",
|
||||
"plt.show()"
|
||||
]
|
||||
}
|
||||
],
|
||||
"metadata": {
|
||||
"kernelspec": {
|
||||
"display_name": "Python 3",
|
||||
"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": "3.11.8"
|
||||
}
|
||||
},
|
||||
"nbformat": 4,
|
||||
"nbformat_minor": 2
|
||||
}
|
||||
@@ -1,61 +0,0 @@
|
||||
|
||||
from torch.utils.data import Dataset
|
||||
from transformers import BertGenerationEncoder, BertGenerationDecoder, EncoderDecoderModel, BertTokenizer
|
||||
from tqdm import tqdm as progress_bar
|
||||
import torch
|
||||
import matplotlib
|
||||
|
||||
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
|
||||
print(device)
|
||||
|
||||
encoder = BertGenerationEncoder.from_pretrained("google-bert/bert-base-uncased", bos_token_id=101, eos_token_id=102)
|
||||
# add cross attention layers and use BERT's cls token as BOS token and sep token as EOS token
|
||||
decoder = BertGenerationDecoder.from_pretrained("google-bert/bert-base-uncased", add_cross_attention=True, is_decoder=True, bos_token_id=101, eos_token_id=102)
|
||||
model = EncoderDecoderModel(encoder=encoder, decoder=decoder)
|
||||
|
||||
# create tokenizer...
|
||||
tokenizer = BertTokenizer.from_pretrained("google-bert/bert-large-uncased")
|
||||
|
||||
import json
|
||||
|
||||
class CodeDataset(Dataset):
|
||||
def __init__(self):
|
||||
with open("data/conala-train.json") as f:
|
||||
self.data = json.load(f)
|
||||
|
||||
def __len__(self):
|
||||
return len(self.data)
|
||||
|
||||
def __getitem__(self, idx):
|
||||
intent = self.data[idx]["rewritten_intent"] if self.data[idx]["rewritten_intent"] else self.data[idx]["intent"]
|
||||
return intent, self.data[idx]["snippet"]
|
||||
|
||||
|
||||
optimizer = torch.optim.AdamW(params=model.parameters(), lr=1e-3)
|
||||
dataloader = CodeDataset()
|
||||
model = model.to(device)
|
||||
|
||||
losses = []
|
||||
epochs = 10
|
||||
for i in range(epochs):
|
||||
|
||||
epoch_loss = 0
|
||||
|
||||
for idx, (question, answer) in progress_bar(enumerate(dataloader), total=len(dataloader)):
|
||||
|
||||
input_ids = tokenizer(question, add_special_tokens=False, return_tensors="pt").input_ids.to(device)
|
||||
label_ids = tokenizer(answer, return_tensors="pt").input_ids.to(device)
|
||||
|
||||
loss = model(input_ids=input_ids, decoder_input_ids=label_ids, labels=label_ids).loss
|
||||
|
||||
optimizer.zero_grad()
|
||||
loss.backward()
|
||||
optimizer.step()
|
||||
|
||||
epoch_loss += loss.item()
|
||||
|
||||
losses.append(epoch_loss)
|
||||
|
||||
plt.plot(losses, color="green", label="Training Loss")
|
||||
plt.legend(loc = 'upper left')
|
||||
plt.savefig("plot.png")
|
||||
Reference in new issue
Block a user