|
import argparse
|
|
import math
|
|
import os
|
|
import torch
|
|
import torch.nn as nn
|
|
import torch.nn.functional as F
|
|
import torch.optim as optim
|
|
from torch.utils.data import DataLoader
|
|
import copy
|
|
from torch.optim.lr_scheduler import CosineAnnealingLR
|
|
from torch.cuda.amp import autocast, GradScaler
|
|
from datasets import load_dataset
|
|
from transformers import AutoTokenizer
|
|
|
|
device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
|
|
|
|
def parse_args():
|
|
parser = argparse.ArgumentParser(description='Train or Inference with World Model and Tree of Thought.')
|
|
parser.add_argument('--model_name', type=str, default='gpt2', help='Pretrained model name or path')
|
|
parser.add_argument('--dataset_name', type=str, default='wikitext', help='Dataset name from HuggingFace Datasets')
|
|
parser.add_argument('--dataset_config', type=str, default='wikitext-2-raw-v1', help='Dataset configuration name')
|
|
parser.add_argument('--batch_size', type=int, default=4, help='Batch size')
|
|
parser.add_argument('--num_epochs', type=int, default=3, help='Number of epochs')
|
|
parser.add_argument('--max_length', type=int, default=128, help='Maximum sequence length')
|
|
parser.add_argument('--mcts_iterations', type=int, default=3, help='Number of MCTS Iterations')
|
|
parser.add_argument('--mcts_exploration_constant', type=float, default=1.414, help='Exploration constant for MCTS')
|
|
parser.add_argument('--accumulation_steps', type=int, default=4, help='Gradient accumulation steps')
|
|
parser.add_argument('--learning_rate', type=float, default=1e-4, help='Learning rate')
|
|
parser.add_argument('--weight_decay', type=float, default=1e-2, help='Weight decay')
|
|
parser.add_argument('--alpha', type=float, default=0.1, help='Entropy regularization weight')
|
|
parser.add_argument('--beta', type=float, default=0.1, help='Variance regularization weight')
|
|
parser.add_argument('--max_grad_norm', type=float, default=1.0, help='Max gradient norm for clipping')
|
|
parser.add_argument('--save_dir', type=str, default='./models', help='Directory to save the models')
|
|
parser.add_argument('--temperature', type=float, default=1.0, help='Temperature parameter for entropy and variance')
|
|
parser.add_argument('--mode', type=str, choices=['train', 'inference'], default='train', help='Mode: train or inference')
|
|
parser.add_argument('--inference_mode', type=str, choices=['world_model', 'without_world_model', 'world_model_tree_of_thought'], default='world_model_tree_of_thought', help='Inference mode')
|
|
parser.add_argument('--query', type=str, default='', help='Input query for inference')
|
|
parser.add_argument('--train_mode', type=str, choices=['world_model', 'language_model'], default='world_model', help='Train world model or language model only')
|
|
|
|
|
|
args, unknown = parser.parse_known_args()
|
|
return args
|
|
|
|
def load_data(args, tokenizer):
|
|
|
|
dataset = load_dataset(args.dataset_name, args.dataset_config)
|
|
|
|
|
|
if tokenizer.pad_token is None:
|
|
tokenizer.pad_token = tokenizer.eos_token
|
|
|
|
def tokenize_function(examples):
|
|
return tokenizer(examples['text'], truncation=True, max_length=args.max_length)
|
|
|
|
tokenized_datasets = dataset.map(
|
|
tokenize_function,
|
|
batched=True,
|
|
num_proc=4,
|
|
remove_columns=dataset['train'].column_names,
|
|
)
|
|
|
|
|
|
block_size = args.max_length
|
|
|
|
def group_texts(examples):
|
|
|
|
concatenated_examples = {k: sum(examples[k], []) for k in examples.keys()}
|
|
total_length = len(concatenated_examples['input_ids'])
|
|
|
|
total_length = (total_length // block_size) * block_size
|
|
|
|
result = {
|
|
k: [t[i : i + block_size] for i in range(0, total_length, block_size)]
|
|
for k, t in concatenated_examples.items()
|
|
}
|
|
result['labels'] = result['input_ids'].copy()
|
|
return result
|
|
|
|
lm_datasets = tokenized_datasets.map(
|
|
group_texts,
|
|
batched=True,
|
|
num_proc=4,
|
|
)
|
|
|
|
|
|
train_dataset = lm_datasets['train']
|
|
eval_dataset = lm_datasets['validation'] if 'validation' in lm_datasets else lm_datasets['test']
|
|
|
|
def data_collator(data):
|
|
return {
|
|
'input_ids': torch.tensor([f['input_ids'] for f in data], dtype=torch.long),
|
|
'labels': torch.tensor([f['labels'] for f in data], dtype=torch.long)
|
|
}
|
|
|
|
train_loader = DataLoader(
|
|
train_dataset,
|
|
shuffle=True,
|
|
batch_size=args.batch_size,
|
|
collate_fn=data_collator,
|
|
pin_memory=True,
|
|
num_workers=4
|
|
)
|
|
eval_loader = DataLoader(
|
|
eval_dataset,
|
|
shuffle=False,
|
|
batch_size=args.batch_size,
|
|
collate_fn=data_collator,
|
|
pin_memory=True,
|
|
num_workers=4
|
|
)
|
|
|
|
return train_loader, eval_loader
|
|
|
|
def save_all_models(transformer_model, representation_network, dynamics_network, prediction_network, action_encoder, save_dir, epoch):
|
|
"""
|
|
Save all models to the specified directory.
|
|
|
|
Args:
|
|
transformer_model (nn.Module): Transformer model.
|
|
representation_network (nn.Module): Representation network.
|
|
dynamics_network (nn.Module): Dynamics network.
|
|
prediction_network (nn.Module): Prediction network.
|
|
action_encoder (nn.Module): Action encoder.
|
|
save_dir (str): Directory to save the models.
|
|
epoch (int): Current epoch number.
|
|
"""
|
|
os.makedirs(save_dir, exist_ok=True)
|
|
|
|
torch.save(transformer_model.state_dict(), os.path.join(save_dir, f'transformer_model_epoch_{epoch}.pt'))
|
|
torch.save(representation_network.state_dict(), os.path.join(save_dir, f'representation_network_epoch_{epoch}.pt'))
|
|
torch.save(dynamics_network.state_dict(), os.path.join(save_dir, f'dynamics_network_epoch_{epoch}.pt'))
|
|
torch.save(prediction_network.state_dict(), os.path.join(save_dir, f'prediction_network_epoch_{epoch}.pt'))
|
|
torch.save(action_encoder.state_dict(), os.path.join(save_dir, f'action_encoder_epoch_{epoch}.pt'))
|
|
|
|
print(f"All models saved for epoch {epoch}.")
|
|
|
|
class RotaryPositionalEncoding(nn.Module):
|
|
def __init__(self, d_model):
|
|
super(RotaryPositionalEncoding, self).__init__()
|
|
inv_freq = 1.0 / (10000 ** (torch.arange(0, d_model, 2).float() / d_model))
|
|
self.register_buffer('inv_freq', inv_freq)
|
|
|
|
def forward(self, x):
|
|
seq_len, batch_size, _ = x.size()
|
|
t = torch.arange(seq_len, device=x.device).type_as(self.inv_freq)
|
|
sinusoid_inp = torch.einsum("i,j->ij", t, self.inv_freq)
|
|
sin = sinusoid_inp.sin().unsqueeze(1)
|
|
cos = sinusoid_inp.cos().unsqueeze(1)
|
|
|
|
x1 = x[..., 0::2]
|
|
x2 = x[..., 1::2]
|
|
|
|
|
|
x_rotated = torch.zeros_like(x)
|
|
x_rotated[..., 0::2] = x1 * cos - x2 * sin
|
|
x_rotated[..., 1::2] = x1 * sin + x2 * cos
|
|
|
|
return x_rotated
|
|
|
|
class MultiHeadAttention(nn.Module):
|
|
def __init__(self, d_model, num_heads):
|
|
super(MultiHeadAttention, self).__init__()
|
|
assert d_model % num_heads == 0, "d_model must be divisible by num_heads"
|
|
self.d_k = d_model // num_heads
|
|
self.num_heads = num_heads
|
|
self.linear_q = nn.Linear(d_model, d_model)
|
|
self.linear_k = nn.Linear(d_model, d_model)
|
|
self.linear_v = nn.Linear(d_model, d_model)
|
|
self.linear_out = nn.Linear(d_model, d_model)
|
|
|
|
def forward(self, query, key, value, mask=None):
|
|
batch_size = query.size(0)
|
|
query = self.linear_q(query).view(batch_size, -1, self.num_heads, self.d_k).transpose(1, 2)
|
|
key = self.linear_k(key).view(batch_size, -1, self.num_heads, self.d_k).transpose(1, 2)
|
|
value = self.linear_v(value).view(batch_size, -1, self.num_heads, self.d_k).transpose(1, 2)
|
|
|
|
scores = torch.matmul(query, key.transpose(-2, -1)) / math.sqrt(self.d_k)
|
|
if mask is not None:
|
|
scores = scores.masked_fill(mask == 0, -1e4)
|
|
attn = F.softmax(scores, dim=-1)
|
|
output = torch.matmul(attn, value)
|
|
|
|
output = output.transpose(1, 2).contiguous().view(batch_size, -1, self.num_heads * self.d_k)
|
|
return self.linear_out(output)
|
|
|
|
class MoE(nn.Module):
|
|
def __init__(self, d_model, num_experts, d_ff, top_k=2, dropout=0.1):
|
|
super(MoE, self).__init__()
|
|
self.num_experts = num_experts
|
|
self.top_k = top_k
|
|
self.experts = nn.ModuleList([
|
|
nn.Sequential(
|
|
nn.Linear(d_model, d_ff),
|
|
nn.GELU() if i % 2 == 0 else nn.SiLU(),
|
|
nn.Linear(d_ff, d_model)
|
|
)
|
|
for i in range(num_experts)
|
|
])
|
|
self.gate = nn.Linear(d_model, num_experts)
|
|
self.dropout = nn.Dropout(dropout)
|
|
|
|
def forward(self, x):
|
|
batch_size, seq_len, d_model = x.size()
|
|
|
|
gate_scores = self.gate(x)
|
|
top_k_scores, top_k_indices = torch.topk(gate_scores, self.top_k, dim=-1)
|
|
top_k_scores = F.softmax(top_k_scores, dim=-1)
|
|
|
|
|
|
output = torch.zeros_like(x)
|
|
|
|
|
|
x_flat = x.view(-1, d_model)
|
|
output_flat = output.view(-1, d_model)
|
|
top_k_indices_flat = top_k_indices.view(-1, self.top_k)
|
|
top_k_scores_flat = top_k_scores.view(-1, self.top_k)
|
|
|
|
for k in range(self.top_k):
|
|
expert_idx_flat = top_k_indices_flat[:, k]
|
|
expert_scores_flat = top_k_scores_flat[:, k]
|
|
for e in range(self.num_experts):
|
|
mask = (expert_idx_flat == e)
|
|
if mask.any():
|
|
x_masked = x_flat[mask]
|
|
expert_output = self.experts[e](x_masked)
|
|
output_flat[mask] += expert_scores_flat[mask].unsqueeze(-1) * expert_output
|
|
|
|
output = output_flat.view(batch_size, seq_len, d_model)
|
|
return self.dropout(output)
|
|
|
|
class TransformerBlock(nn.Module):
|
|
def __init__(self, d_model, num_heads, d_ff, num_experts, dropout=0.1, top_k=2):
|
|
super(TransformerBlock, self).__init__()
|
|
self.self_attention = MultiHeadAttention(d_model, num_heads)
|
|
self.norm1 = nn.LayerNorm(d_model)
|
|
self.cross_attention = MultiHeadAttention(d_model, num_heads)
|
|
self.norm2 = nn.LayerNorm(d_model)
|
|
self.moe = MoE(d_model, num_experts, d_ff, top_k, dropout)
|
|
self.norm3 = nn.LayerNorm(d_model)
|
|
|
|
def forward(self, x, mask=None, enc_output=None, enc_mask=None):
|
|
|
|
attn_output = self.self_attention(x, x, x, mask)
|
|
x = self.norm1(x + attn_output)
|
|
|
|
if enc_output is not None:
|
|
cross_attn_output = self.cross_attention(x, enc_output, enc_output, enc_mask)
|
|
x = self.norm2(x + cross_attn_output)
|
|
|
|
moe_output = self.moe(x)
|
|
return self.norm3(x + moe_output)
|
|
|
|
class Transformer(nn.Module):
|
|
def __init__(self, input_dim, d_model, num_heads, num_layers, d_ff, num_experts, output_dim, dropout=0.1, top_k=2):
|
|
super(Transformer, self).__init__()
|
|
self.embedding = nn.Embedding(input_dim, d_model, padding_idx=input_dim - 1)
|
|
self.rotary_positional_encoding = RotaryPositionalEncoding(d_model)
|
|
self.encoder_layers = nn.ModuleList(
|
|
[TransformerBlock(d_model, num_heads, d_ff, num_experts, dropout, top_k) for _ in range(num_layers)]
|
|
)
|
|
self.decoder_layers = nn.ModuleList(
|
|
[TransformerBlock(d_model, num_heads, d_ff, num_experts, dropout, top_k) for _ in range(num_layers)]
|
|
)
|
|
self.output_layer = nn.Linear(d_model, output_dim)
|
|
self.d_model = d_model
|
|
|
|
def forward(self, src, tgt, src_mask=None, tgt_mask=None):
|
|
|
|
src = self.embedding(src) * math.sqrt(self.d_model)
|
|
src = src.transpose(0, 1)
|
|
src = self.rotary_positional_encoding(src)
|
|
src = src.transpose(0, 1)
|
|
for layer in self.encoder_layers:
|
|
src = layer(src, src_mask)
|
|
|
|
|
|
tgt = self.embedding(tgt) * math.sqrt(self.d_model)
|
|
tgt = tgt.transpose(0, 1)
|
|
tgt = self.rotary_positional_encoding(tgt)
|
|
tgt = tgt.transpose(0, 1)
|
|
for layer in self.decoder_layers:
|
|
tgt = layer(tgt, tgt_mask, src, src_mask)
|
|
output = self.output_layer(tgt)
|
|
return output
|
|
|
|
def generate(self, src, tokenizer, max_length=20, temperature=1.0):
|
|
"""
|
|
Generate sequences using differentiable sampling (Gumbel-Softmax).
|
|
|
|
Args:
|
|
src (torch.Tensor): Source input tensor of shape (batch_size, seq_len)
|
|
tokenizer (transformers.PreTrainedTokenizer): Tokenizer to access special tokens
|
|
max_length (int): Maximum length of the generated sequence
|
|
temperature (float): Temperature parameter for Gumbel-Softmax
|
|
|
|
Returns:
|
|
torch.Tensor: Generated sequences of shape (batch_size, max_length)
|
|
torch.Tensor: Entropy values for each time step
|
|
torch.Tensor: Variance values for each time step
|
|
"""
|
|
batch_size = src.size(0)
|
|
|
|
|
|
src_enc = self.embedding(src) * math.sqrt(self.d_model)
|
|
src_enc = src_enc.transpose(0, 1)
|
|
src_enc = self.rotary_positional_encoding(src_enc)
|
|
src_enc = src_enc.transpose(0, 1)
|
|
for layer in self.encoder_layers:
|
|
src_enc = layer(src_enc)
|
|
|
|
|
|
tgt_seq = torch.full((batch_size, 1), tokenizer.bos_token_id, dtype=torch.long, device=src.device)
|
|
entropies = []
|
|
variances = []
|
|
|
|
for _ in range(max_length):
|
|
tgt_emb = self.embedding(tgt_seq) * math.sqrt(self.d_model)
|
|
tgt_emb = tgt_emb.transpose(0, 1)
|
|
tgt_emb = self.rotary_positional_encoding(tgt_emb)
|
|
tgt_emb = tgt_emb.transpose(0, 1)
|
|
tgt_dec = tgt_emb
|
|
for layer in self.decoder_layers:
|
|
tgt_dec = layer(tgt_dec, None, src_enc, None)
|
|
output = self.output_layer(tgt_dec)
|
|
logits = output[:, -1, :]
|
|
|
|
|
|
probs = F.softmax(logits / temperature, dim=-1)
|
|
|
|
|
|
entropy = -torch.sum(probs * torch.log(probs + 1e-9), dim=-1)
|
|
entropies.append(entropy)
|
|
|
|
|
|
gumbel_noise = -torch.log(-torch.log(torch.rand_like(probs) + 1e-9) + 1e-9)
|
|
y = (logits + gumbel_noise) / temperature
|
|
y = F.softmax(y, dim=-1)
|
|
|
|
|
|
variance = torch.var(y, dim=-1)
|
|
variances.append(variance)
|
|
|
|
|
|
next_tokens = torch.argmax(y, dim=-1, keepdim=True)
|
|
tgt_seq = torch.cat([tgt_seq, next_tokens], dim=1)
|
|
|
|
|
|
entropies = torch.stack(entropies, dim=1)
|
|
variances = torch.stack(variances, dim=1)
|
|
|
|
return tgt_seq[:, 1:], entropies, variances
|
|
|
|
|
|
|
|
class InfoNCE_Loss(nn.Module):
|
|
def __init__(self, temperature=0.07):
|
|
super(InfoNCE_Loss, self).__init__()
|
|
self.temperature = temperature
|
|
self.cross_entropy = nn.CrossEntropyLoss()
|
|
|
|
def forward(self, z_i, z_j):
|
|
"""
|
|
Args:
|
|
z_i (torch.Tensor): Flattened representations from view i, shape (2n, embed_dim)
|
|
z_j (torch.Tensor): Flattened representations from view j, shape (2n, embed_dim)
|
|
|
|
Returns:
|
|
torch.Tensor: InfoNCE loss
|
|
"""
|
|
n = z_i.size(0)
|
|
z = torch.cat([z_i, z_j], dim=0)
|
|
|
|
z = F.normalize(z, dim=1)
|
|
similarity_matrix = torch.matmul(z, z.T)
|
|
|
|
|
|
mask = torch.eye(2 * n, device=z.device, dtype=torch.bool)
|
|
similarity_matrix = similarity_matrix.masked_fill(mask, -1e4)
|
|
|
|
|
|
labels = torch.arange(n, device=z.device)
|
|
labels = torch.cat([labels + n, labels], dim=0)
|
|
|
|
|
|
similarity_matrix /= self.temperature
|
|
|
|
|
|
loss = self.cross_entropy(similarity_matrix, labels)
|
|
return loss
|
|
|
|
class CovarianceRegularization(nn.Module):
|
|
def __init__(self, lambda_reg=1e-3):
|
|
super(CovarianceRegularization, self).__init__()
|
|
self.lambda_reg = lambda_reg
|
|
|
|
def forward(self, embeddings):
|
|
"""
|
|
Args:
|
|
embeddings (torch.Tensor): Embedding tensor, shape (batch_size, embed_dim)
|
|
|
|
Returns:
|
|
torch.Tensor: Covariance regularization loss
|
|
"""
|
|
batch_size, embed_dim = embeddings.size()
|
|
mean = embeddings.mean(dim=0)
|
|
embeddings_centered = embeddings - mean
|
|
cov = (embeddings_centered.T @ embeddings_centered) / (batch_size - 1)
|
|
cov_loss = torch.sum(cov ** 2) - torch.sum(torch.diag(cov) ** 2)
|
|
return self.lambda_reg * cov_loss
|
|
|
|
class DynamicsPerformanceLoss(nn.Module):
|
|
def __init__(self, lambda_var=1e-3):
|
|
super(DynamicsPerformanceLoss, self).__init__()
|
|
self.lambda_var = lambda_var
|
|
|
|
def forward(self, true_next_state, predicted_next_state):
|
|
"""
|
|
Args:
|
|
true_next_state (torch.Tensor): Ground truth next state, shape (batch_size, state_dim)
|
|
predicted_next_state (torch.Tensor): Predicted next state, shape (batch_size, state_dim)
|
|
|
|
Returns:
|
|
torch.Tensor: Dynamics performance loss
|
|
"""
|
|
mse_loss = F.mse_loss(predicted_next_state, true_next_state)
|
|
variance_loss = torch.var(predicted_next_state, dim=0).mean()
|
|
return mse_loss + self.lambda_var * variance_loss
|
|
|
|
class ThoughtConsistencyLoss(nn.Module):
|
|
def __init__(self):
|
|
super(ThoughtConsistencyLoss, self).__init__()
|
|
|
|
def forward(self, true_next_state, perturbed_next_state):
|
|
"""
|
|
Args:
|
|
true_next_state (torch.Tensor): Ground truth next state, shape (batch_size, state_dim)
|
|
perturbed_next_state (torch.Tensor): Perturbed next state, shape (batch_size, state_dim)
|
|
|
|
Returns:
|
|
torch.Tensor: Thought-consistency loss
|
|
"""
|
|
return F.mse_loss(true_next_state, perturbed_next_state)
|
|
|
|
class PolicyValueJointLoss(nn.Module):
|
|
def __init__(self, lambda_value=0.5):
|
|
super(PolicyValueJointLoss, self).__init__()
|
|
self.lambda_value = lambda_value
|
|
self.cross_entropy = nn.CrossEntropyLoss()
|
|
self.mse_loss = nn.MSELoss()
|
|
|
|
def forward(self, policy_logits, true_policy, value_pred, true_value):
|
|
"""
|
|
Args:
|
|
policy_logits (torch.Tensor): Logits from the policy network, shape (batch_size * seq_len, num_actions)
|
|
true_policy (torch.Tensor): Ground truth policy, shape (batch_size * seq_len, num_actions)
|
|
value_pred (torch.Tensor): Predicted values, shape (batch_size * seq_len)
|
|
true_value (torch.Tensor): Ground truth values, shape (batch_size * seq_len)
|
|
|
|
Returns:
|
|
torch.Tensor: Combined policy and value loss
|
|
"""
|
|
policy_logits = policy_logits.view(-1, policy_logits.size(-1))
|
|
true_policy = true_policy.view(-1, true_policy.size(-1))
|
|
value_pred = value_pred.view(-1)
|
|
true_value = true_value.view(-1)
|
|
|
|
policy_loss = self.cross_entropy(policy_logits, true_policy.argmax(dim=1))
|
|
value_loss = self.mse_loss(value_pred, true_value)
|
|
return policy_loss + self.lambda_value * value_loss
|
|
|
|
class ActionDiversityReward(nn.Module):
|
|
def __init__(self, lambda_div=1e-3):
|
|
super(ActionDiversityReward, self).__init__()
|
|
self.lambda_div = lambda_div
|
|
|
|
def forward(self, action_embeddings):
|
|
"""
|
|
Args:
|
|
action_embeddings (torch.Tensor): Embeddings of actions, shape (batch_size, embed_dim)
|
|
|
|
Returns:
|
|
torch.Tensor: Action diversity loss
|
|
"""
|
|
similarity_matrix = F.cosine_similarity(action_embeddings.unsqueeze(1), action_embeddings.unsqueeze(0), dim=2)
|
|
|
|
similarity_matrix = similarity_matrix - torch.eye(similarity_matrix.size(0)).to(action_embeddings.device)
|
|
diversity_loss = torch.sum(similarity_matrix ** 2)
|
|
return self.lambda_div * diversity_loss
|
|
|
|
class ExpectedThoughtValueLoss(nn.Module):
|
|
def __init__(self):
|
|
super(ExpectedThoughtValueLoss, self).__init__()
|
|
|
|
def forward(self, mcts_best_values):
|
|
"""
|
|
Args:
|
|
mcts_best_values (torch.Tensor): Best values from MCTS, shape (batch_size)
|
|
|
|
Returns:
|
|
torch.Tensor: ETV loss
|
|
"""
|
|
return -mcts_best_values.mean()
|
|
|
|
class ExplorationRegularization(nn.Module):
|
|
def __init__(self, lambda_expl=1e-3):
|
|
super(ExplorationRegularization, self).__init__()
|
|
self.lambda_expl = lambda_expl
|
|
|
|
def forward(self, visit_counts):
|
|
"""
|
|
Args:
|
|
visit_counts (torch.Tensor): Visit counts for actions, shape (batch_size, num_actions)
|
|
|
|
Returns:
|
|
torch.Tensor: Exploration regularization loss
|
|
"""
|
|
reward = torch.sum(1.0 / (visit_counts + 1), dim=-1)
|
|
return self.lambda_expl * reward.mean()
|
|
|
|
class KL_DivergenceLoss(nn.Module):
|
|
def __init__(self):
|
|
super(KL_DivergenceLoss, self).__init__()
|
|
|
|
def forward(self, old_policy, new_policy):
|
|
"""
|
|
Args:
|
|
old_policy (torch.Tensor): Old policy probabilities, shape (batch_size, num_actions)
|
|
new_policy (torch.Tensor): New policy probabilities, shape (batch_size, num_actions)
|
|
|
|
Returns:
|
|
torch.Tensor: KL divergence loss
|
|
"""
|
|
kl_div = F.kl_div(new_policy.log(), old_policy, reduction='batchmean')
|
|
return kl_div
|
|
|
|
|
|
|
|
class ActionEncoder(nn.Module):
|
|
def __init__(self, action_vocab_size, embed_dim):
|
|
super(ActionEncoder, self).__init__()
|
|
self.embedding = nn.Embedding(action_vocab_size, embed_dim)
|
|
|
|
def forward(self, action_indices):
|
|
"""
|
|
Args:
|
|
action_indices (torch.Tensor): Tensor of shape (batch_size, seq_len)
|
|
|
|
Returns:
|
|
torch.Tensor: Encoded actions of shape (batch_size, seq_len, embed_dim)
|
|
"""
|
|
return self.embedding(action_indices)
|
|
|
|
class RepresentationNetwork(nn.Module):
|
|
def __init__(self, vocab_dim, d_model, state_dim):
|
|
super(RepresentationNetwork, self).__init__()
|
|
self.proj = nn.Linear(vocab_dim, d_model)
|
|
self.linear = nn.Linear(d_model, state_dim)
|
|
self.norm = nn.LayerNorm(state_dim)
|
|
|
|
def forward(self, transformer_output):
|
|
"""
|
|
Args:
|
|
transformer_output (torch.Tensor): Shape (batch_size, seq_len, vocab_dim)
|
|
|
|
Returns:
|
|
torch.Tensor: Encoded state of shape (batch_size, seq_len, state_dim)
|
|
"""
|
|
|
|
projected_output = self.proj(transformer_output)
|
|
|
|
state = self.linear(projected_output)
|
|
state = self.norm(state)
|
|
return state
|
|
|
|
class DynamicsNetwork(nn.Module):
|
|
def __init__(self, state_dim, action_dim, hidden_dim):
|
|
super(DynamicsNetwork, self).__init__()
|
|
self.rms_norm = nn.LayerNorm(state_dim)
|
|
self.fc1 = nn.Linear(state_dim + action_dim, hidden_dim)
|
|
self.activation = nn.GELU()
|
|
self.fc2 = nn.Linear(hidden_dim, state_dim)
|
|
|
|
def forward(self, state, action):
|
|
"""
|
|
Args:
|
|
state (torch.Tensor): Current state, shape (batch_size, seq_len, state_dim)
|
|
action (torch.Tensor): Action embedding, shape (batch_size, seq_len, action_dim)
|
|
|
|
Returns:
|
|
torch.Tensor: Predicted next state, shape (batch_size, seq_len, state_dim)
|
|
"""
|
|
norm_state = self.rms_norm(state)
|
|
combined = torch.cat([norm_state, action], dim=-1)
|
|
hidden = self.activation(self.fc1(combined))
|
|
next_state = self.fc2(hidden)
|
|
return next_state
|
|
|
|
class PredictionNetwork(nn.Module):
|
|
def __init__(self, state_dim, action_vocab_size, value_dim):
|
|
super(PredictionNetwork, self).__init__()
|
|
self.state_dim = state_dim
|
|
self.rms_norm = nn.LayerNorm(state_dim)
|
|
self.policy_head = nn.Linear(state_dim, action_vocab_size)
|
|
self.value_head = nn.Linear(state_dim, value_dim)
|
|
|
|
def forward(self, state):
|
|
"""
|
|
Args:
|
|
state (torch.Tensor): State representation, shape (batch_size, seq_len, state_dim)
|
|
Returns:
|
|
Tuple[torch.Tensor, torch.Tensor]: Policy logits and value estimates
|
|
"""
|
|
norm_state = self.rms_norm(state)
|
|
policy_logits = self.policy_head(norm_state)
|
|
value_estimates = self.value_head(norm_state).squeeze(-1)
|
|
return policy_logits, value_estimates
|
|
|
|
|
|
|
|
|
|
class ThoughtNode:
|
|
def __init__(self, name):
|
|
self.name = name
|
|
self.children = []
|
|
self.parent = None
|
|
|
|
def add_child(self, child_node):
|
|
child_node.parent = self
|
|
self.children.append(child_node)
|
|
|
|
|
|
def build_tree_of_thought():
|
|
|
|
root = ThoughtNode('Problem-Solving Process')
|
|
|
|
|
|
problem_identification = ThoughtNode('Problem Identification')
|
|
problem_analysis = ThoughtNode('Problem Analysis')
|
|
solution_generation = ThoughtNode('Solution Generation')
|
|
implementation = ThoughtNode('Implementation')
|
|
evaluation_adjustment = ThoughtNode('Evaluation and Adjustment')
|
|
|
|
root.add_child(problem_identification)
|
|
root.add_child(problem_analysis)
|
|
root.add_child(solution_generation)
|
|
root.add_child(implementation)
|
|
root.add_child(evaluation_adjustment)
|
|
|
|
|
|
B1 = ThoughtNode('Define the Problem')
|
|
B2 = ThoughtNode('Identify Stakeholders')
|
|
B3 = ThoughtNode('Determine Constraints')
|
|
B4 = ThoughtNode('Recognize Problem Type')
|
|
B5 = ThoughtNode('Historical Context')
|
|
problem_identification.add_child(B1)
|
|
problem_identification.add_child(B2)
|
|
problem_identification.add_child(B3)
|
|
problem_identification.add_child(B4)
|
|
problem_identification.add_child(B5)
|
|
|
|
|
|
B1a = ThoughtNode('Problem Statement Formulation')
|
|
B1b = ThoughtNode('Scope Definition')
|
|
B1c = ThoughtNode('Objective Setting')
|
|
B1.add_child(B1a)
|
|
B1.add_child(B1b)
|
|
B1.add_child(B1c)
|
|
|
|
|
|
B2a = ThoughtNode('Stakeholder Mapping')
|
|
B2b = ThoughtNode('Interest and Influence Analysis')
|
|
B2c = ThoughtNode('Engagement Strategy')
|
|
B2.add_child(B2a)
|
|
B2.add_child(B2b)
|
|
B2.add_child(B2c)
|
|
|
|
|
|
B3a = ThoughtNode('Resource Limitations')
|
|
B3b = ThoughtNode('Time Constraints')
|
|
B3c = ThoughtNode('Legal and Regulatory Constraints')
|
|
B3.add_child(B3a)
|
|
B3.add_child(B3b)
|
|
B3.add_child(B3c)
|
|
|
|
|
|
B4a = ThoughtNode('Simple vs Complex')
|
|
B4b = ThoughtNode('Known vs Unknown')
|
|
B4c = ThoughtNode('Tame vs Wicked Problems')
|
|
B4.add_child(B4a)
|
|
B4.add_child(B4b)
|
|
B4.add_child(B4c)
|
|
|
|
|
|
B5a = ThoughtNode('Previous Attempts')
|
|
B5b = ThoughtNode('Lessons Learned')
|
|
B5c = ThoughtNode('Environmental Factors')
|
|
B5.add_child(B5a)
|
|
B5.add_child(B5b)
|
|
B5.add_child(B5c)
|
|
|
|
|
|
C1 = ThoughtNode('Root Cause Analysis')
|
|
C2 = ThoughtNode('System Mapping')
|
|
C3 = ThoughtNode('Data Collection')
|
|
C4 = ThoughtNode('Impact Assessment')
|
|
C5 = ThoughtNode('Theoretical Framework')
|
|
problem_analysis.add_child(C1)
|
|
problem_analysis.add_child(C2)
|
|
problem_analysis.add_child(C3)
|
|
problem_analysis.add_child(C4)
|
|
problem_analysis.add_child(C5)
|
|
|
|
|
|
C1a = ThoughtNode('5 Whys Technique')
|
|
C1b = ThoughtNode('Fishbone Diagram')
|
|
C1c = ThoughtNode('Pareto Analysis')
|
|
C1.add_child(C1a)
|
|
C1.add_child(C1b)
|
|
C1.add_child(C1c)
|
|
|
|
|
|
C2a = ThoughtNode('Causal Loop Diagrams')
|
|
C2b = ThoughtNode('Stock and Flow Models')
|
|
C2c = ThoughtNode('Network Analysis')
|
|
C2.add_child(C2a)
|
|
C2.add_child(C2b)
|
|
C2.add_child(C2c)
|
|
|
|
|
|
C3a = ThoughtNode('Quantitative Data')
|
|
C3b = ThoughtNode('Qualitative Data')
|
|
C3c = ThoughtNode('Data Validation')
|
|
C3.add_child(C3a)
|
|
C3.add_child(C3b)
|
|
C3.add_child(C3c)
|
|
|
|
|
|
C3a1 = ThoughtNode('Surveys and Questionnaires')
|
|
C3a2 = ThoughtNode('Experimental Data')
|
|
C3a3 = ThoughtNode('Big Data Analytics')
|
|
C3a.add_child(C3a1)
|
|
C3a.add_child(C3a2)
|
|
C3a.add_child(C3a3)
|
|
|
|
|
|
C3b1 = ThoughtNode('Interviews')
|
|
C3b2 = ThoughtNode('Focus Groups')
|
|
C3b3 = ThoughtNode('Observational Studies')
|
|
C3b.add_child(C3b1)
|
|
C3b.add_child(C3b2)
|
|
C3b.add_child(C3b3)
|
|
|
|
|
|
C3c1 = ThoughtNode('Statistical Validation')
|
|
C3c2 = ThoughtNode('Cross-Validation')
|
|
C3c3 = ThoughtNode('Expert Review')
|
|
C3c.add_child(C3c1)
|
|
C3c.add_child(C3c2)
|
|
C3c.add_child(C3c3)
|
|
|
|
|
|
C4a = ThoughtNode('Environmental Impact')
|
|
C4b = ThoughtNode('Social Impact')
|
|
C4c = ThoughtNode('Economic Impact')
|
|
C4.add_child(C4a)
|
|
C4.add_child(C4b)
|
|
C4.add_child(C4c)
|
|
|
|
|
|
C5a = ThoughtNode('Literature Review')
|
|
C5b = ThoughtNode('Conceptual Modeling')
|
|
C5c = ThoughtNode('Hypothesis Formation')
|
|
C5.add_child(C5a)
|
|
C5.add_child(C5b)
|
|
C5.add_child(C5c)
|
|
|
|
|
|
D1 = ThoughtNode('Creative Problem Solving')
|
|
D2 = ThoughtNode('Analytical Approach')
|
|
D3 = ThoughtNode('Mathematical Computation')
|
|
D4 = ThoughtNode('Decision Making')
|
|
solution_generation.add_child(D1)
|
|
solution_generation.add_child(D2)
|
|
solution_generation.add_child(D3)
|
|
solution_generation.add_child(D4)
|
|
|
|
|
|
E1 = ThoughtNode('Action Planning')
|
|
E2 = ThoughtNode('Resource Allocation')
|
|
E3 = ThoughtNode('Change Management')
|
|
implementation.add_child(E1)
|
|
implementation.add_child(E2)
|
|
implementation.add_child(E3)
|
|
|
|
|
|
F1 = ThoughtNode('Verification')
|
|
F2 = ThoughtNode('Performance Metrics')
|
|
F3 = ThoughtNode('Feedback Loops')
|
|
F4 = ThoughtNode('Continuous Improvement')
|
|
evaluation_adjustment.add_child(F1)
|
|
evaluation_adjustment.add_child(F2)
|
|
evaluation_adjustment.add_child(F3)
|
|
evaluation_adjustment.add_child(F4)
|
|
|
|
|
|
G = ThoughtNode('Cross-Cutting Considerations')
|
|
root.add_child(G)
|
|
|
|
|
|
G1 = ThoughtNode('Ethical Framework')
|
|
G2 = ThoughtNode('Stakeholder Management')
|
|
G3 = ThoughtNode('Interdisciplinary Connections')
|
|
G4 = ThoughtNode('Technological Integration')
|
|
G5 = ThoughtNode('Emotional Intelligence')
|
|
G6 = ThoughtNode('Collaborative Problem Solving')
|
|
G7 = ThoughtNode('Computational Considerations')
|
|
G8 = ThoughtNode('Order of Operations')
|
|
G9 = ThoughtNode('Critical Thinking')
|
|
G10 = ThoughtNode('Future Perspective')
|
|
G11 = ThoughtNode('Learning and Adaptation')
|
|
G.add_child(G1)
|
|
G.add_child(G2)
|
|
G.add_child(G3)
|
|
G.add_child(G4)
|
|
G.add_child(G5)
|
|
G.add_child(G6)
|
|
G.add_child(G7)
|
|
G.add_child(G8)
|
|
G.add_child(G9)
|
|
G.add_child(G10)
|
|
G.add_child(G11)
|
|
|
|
|
|
G1a = ThoughtNode('Value-based Decision Making')
|
|
G1b = ThoughtNode('Long-term Consequences')
|
|
G1.add_child(G1a)
|
|
G1.add_child(G1b)
|
|
|
|
|
|
G1a1 = ThoughtNode('Ethical Theories Application')
|
|
G1a2 = ThoughtNode('Moral Dilemma Resolution')
|
|
G1a.add_child(G1a1)
|
|
G1a.add_child(G1a2)
|
|
|
|
|
|
G1b1 = ThoughtNode('Sustainability Assessment')
|
|
G1b2 = ThoughtNode('Intergenerational Impact')
|
|
G1b.add_child(G1b1)
|
|
G1b.add_child(G1b2)
|
|
|
|
|
|
G2a = ThoughtNode('Direct Stakeholders')
|
|
G2b = ThoughtNode('Indirect Stakeholders')
|
|
G2c = ThoughtNode('Conflicting Interests')
|
|
G2.add_child(G2a)
|
|
G2.add_child(G2b)
|
|
G2.add_child(G2c)
|
|
|
|
|
|
G2c1 = ThoughtNode('Negotiation Strategies')
|
|
G2c2 = ThoughtNode('Conflict Resolution Techniques')
|
|
G2c.add_child(G2c1)
|
|
G2c.add_child(G2c2)
|
|
|
|
|
|
G3a = ThoughtNode('Related Fields')
|
|
G3b = ThoughtNode('Cross-disciplinary Impact')
|
|
G3.add_child(G3a)
|
|
G3.add_child(G3b)
|
|
|
|
|
|
G3a1 = ThoughtNode('Cross-domain Knowledge Transfer')
|
|
G3a2 = ThoughtNode('Interdisciplinary Collaboration')
|
|
G3a.add_child(G3a1)
|
|
G3a.add_child(G3a2)
|
|
|
|
|
|
G3b1 = ThoughtNode('Synergy Identification')
|
|
G3b2 = ThoughtNode('Holistic Impact Assessment')
|
|
G3b.add_child(G3b1)
|
|
G3b.add_child(G3b2)
|
|
|
|
|
|
G4a = ThoughtNode('AI-assisted Problem Solving')
|
|
G4b = ThoughtNode('Data-driven Insights')
|
|
G4c = ThoughtNode('Digital Collaboration Tools')
|
|
G4.add_child(G4a)
|
|
G4.add_child(G4b)
|
|
G4.add_child(G4c)
|
|
|
|
|
|
G4a1 = ThoughtNode('Machine Learning Models')
|
|
G4a2 = ThoughtNode('Natural Language Processing')
|
|
G4a.add_child(G4a1)
|
|
G4a.add_child(G4a2)
|
|
|
|
|
|
G4b1 = ThoughtNode('Big Data Analytics')
|
|
G4b2 = ThoughtNode('Predictive Modeling')
|
|
G4b.add_child(G4b1)
|
|
G4b.add_child(G4b2)
|
|
|
|
|
|
G4c1 = ThoughtNode('Project Management Platforms')
|
|
G4c2 = ThoughtNode('Virtual Reality Collaboration')
|
|
G4c.add_child(G4c1)
|
|
G4c.add_child(G4c2)
|
|
|
|
|
|
G5a = ThoughtNode('Self-Awareness')
|
|
G5b = ThoughtNode('Empathy')
|
|
G5c = ThoughtNode('Stress Management')
|
|
G5.add_child(G5a)
|
|
G5.add_child(G5b)
|
|
G5.add_child(G5c)
|
|
|
|
|
|
G5a1 = ThoughtNode('Emotional Recognition')
|
|
G5a2 = ThoughtNode('Personal Bias Identification')
|
|
G5a.add_child(G5a1)
|
|
G5a.add_child(G5a2)
|
|
|
|
|
|
G5b1 = ThoughtNode('Perspective Taking')
|
|
G5b2 = ThoughtNode('Active Listening')
|
|
G5b.add_child(G5b1)
|
|
G5b.add_child(G5b2)
|
|
|
|
|
|
G5c1 = ThoughtNode('Mindfulness Techniques')
|
|
G5c2 = ThoughtNode('Resilience Building')
|
|
G5c.add_child(G5c1)
|
|
G5c.add_child(G5c2)
|
|
|
|
|
|
G6a = ThoughtNode('Team Dynamics')
|
|
G6b = ThoughtNode('Communication Strategies')
|
|
G6c = ThoughtNode('Conflict Resolution')
|
|
G6.add_child(G6a)
|
|
G6.add_child(G6b)
|
|
G6.add_child(G6c)
|
|
|
|
|
|
G6a1 = ThoughtNode('Team Formation Strategies')
|
|
G6a2 = ThoughtNode('Role Assignment')
|
|
G6a.add_child(G6a1)
|
|
G6a.add_child(G6a2)
|
|
|
|
|
|
G6b1 = ThoughtNode('Clear Messaging')
|
|
G6b2 = ThoughtNode('Feedback Mechanisms')
|
|
G6b.add_child(G6b1)
|
|
G6b.add_child(G6b2)
|
|
|
|
|
|
G6c1 = ThoughtNode('Mediation Techniques')
|
|
G6c2 = ThoughtNode('Consensus Building')
|
|
G6c.add_child(G6c1)
|
|
G6c.add_child(G6c2)
|
|
|
|
|
|
G7a = ThoughtNode('CPU Operations')
|
|
G7b = ThoughtNode('GPU Parallelization')
|
|
G7c = ThoughtNode('Floating-Point Precision')
|
|
G7.add_child(G7a)
|
|
G7.add_child(G7b)
|
|
G7.add_child(G7c)
|
|
|
|
|
|
G7a1 = ThoughtNode('Instruction Set Architecture')
|
|
G7a2 = ThoughtNode('Pipelining and Parallelism')
|
|
G7a.add_child(G7a1)
|
|
G7a.add_child(G7a2)
|
|
|
|
|
|
G7b1 = ThoughtNode('CUDA Programming')
|
|
G7b2 = ThoughtNode('OpenCL Framework')
|
|
G7b.add_child(G7b1)
|
|
G7b.add_child(G7b2)
|
|
|
|
|
|
G7c1 = ThoughtNode('IEEE 754 Standard')
|
|
G7c2 = ThoughtNode('Error Propagation Analysis')
|
|
G7c.add_child(G7c1)
|
|
G7c.add_child(G7c2)
|
|
|
|
|
|
G8a = ThoughtNode('Parentheses')
|
|
G8b = ThoughtNode('Exponents')
|
|
G8c = ThoughtNode('Multiplication and Division')
|
|
G8d = ThoughtNode('Addition and Subtraction')
|
|
G8.add_child(G8a)
|
|
G8.add_child(G8b)
|
|
G8.add_child(G8c)
|
|
G8.add_child(G8d)
|
|
|
|
|
|
G9a = ThoughtNode('Assumptions Questioning')
|
|
G9b = ThoughtNode('Bias Recognition')
|
|
G9.add_child(G9a)
|
|
G9.add_child(G9b)
|
|
|
|
|
|
G9a1 = ThoughtNode('Socratic Questioning')
|
|
G9a2 = ThoughtNode('Devil\'s Advocate Approach')
|
|
G9a.add_child(G9a1)
|
|
G9a.add_child(G9a2)
|
|
|
|
|
|
G9b1 = ThoughtNode('Cognitive Bias Identification')
|
|
G9b2 = ThoughtNode('Debiasing Techniques')
|
|
G9b.add_child(G9b1)
|
|
G9b.add_child(G9b2)
|
|
|
|
|
|
G10a = ThoughtNode('Short-term Projections')
|
|
G10b = ThoughtNode('Long-term Scenarios')
|
|
G10c = ThoughtNode('Potential Impacts')
|
|
G10.add_child(G10a)
|
|
G10.add_child(G10b)
|
|
G10.add_child(G10c)
|
|
|
|
|
|
G10a1 = ThoughtNode('Trend Analysis')
|
|
G10a2 = ThoughtNode('Scenario Planning')
|
|
G10a.add_child(G10a1)
|
|
G10a.add_child(G10a2)
|
|
|
|
|
|
G10b1 = ThoughtNode('Futures Wheel')
|
|
G10b2 = ThoughtNode('Backcasting')
|
|
G10b.add_child(G10b1)
|
|
G10b.add_child(G10b2)
|
|
|
|
|
|
G10c1 = ThoughtNode('Risk Assessment')
|
|
G10c2 = ThoughtNode('Opportunity Identification')
|
|
G10c.add_child(G10c1)
|
|
G10c.add_child(G10c2)
|
|
|
|
|
|
G11a = ThoughtNode('Reflective Practice')
|
|
G11b = ThoughtNode('Knowledge Transfer')
|
|
G11c = ThoughtNode('Adaptive Problem Solving')
|
|
G11.add_child(G11a)
|
|
G11.add_child(G11b)
|
|
G11.add_child(G11c)
|
|
|
|
|
|
G11a1 = ThoughtNode('After Action Review')
|
|
G11a2 = ThoughtNode('Learning Journals')
|
|
G11a.add_child(G11a1)
|
|
G11a.add_child(G11a2)
|
|
|
|
|
|
G11b1 = ThoughtNode('Best Practice Documentation')
|
|
G11b2 = ThoughtNode('Mentoring Programs')
|
|
G11b.add_child(G11b1)
|
|
G11b.add_child(G11b2)
|
|
|
|
|
|
G11c1 = ThoughtNode('Iterative Approaches')
|
|
G11c2 = ThoughtNode('Flexibility in Methodology')
|
|
G11c.add_child(G11c1)
|
|
G11c.add_child(G11c2)
|
|
|
|
return root
|
|
|
|
def traverse_tree(node, action_list):
|
|
if node.name not in action_list:
|
|
action_list.append(node.name)
|
|
for child in node.children:
|
|
traverse_tree(child, action_list)
|
|
|
|
class MCTSNode:
|
|
__slots__ = [
|
|
'state',
|
|
'parent',
|
|
'action',
|
|
'children',
|
|
'visit_count',
|
|
'value_sum',
|
|
'prior',
|
|
'cached_policy',
|
|
'cached_value',
|
|
'thought_node'
|
|
]
|
|
|
|
def __init__(self, state, thought_node, parent=None, action=None):
|
|
self.state = state
|
|
self.thought_node = thought_node
|
|
self.parent = parent
|
|
self.action = action
|
|
self.children = {}
|
|
self.visit_count = 0
|
|
self.value_sum = 0.0
|
|
self.prior = 0.0
|
|
self.cached_policy = None
|
|
self.cached_value = None
|
|
|
|
def expand(self, priors):
|
|
"""
|
|
Expand the node by adding all valid child nodes from the thought tree.
|
|
|
|
Args:
|
|
priors (dict): A dictionary mapping action names to prior probabilities.
|
|
"""
|
|
for child_thought_node in self.thought_node.children:
|
|
action = child_thought_node.name
|
|
if action not in self.children:
|
|
|
|
child_state = self.state.apply_action(action)
|
|
child_node = MCTSNode(
|
|
state=child_state,
|
|
thought_node=child_thought_node,
|
|
parent=self,
|
|
action=action
|
|
)
|
|
child_node.prior = priors.get(action, 1.0 / len(self.thought_node.children))
|
|
self.children[action] = child_node
|
|
|
|
def is_leaf(self):
|
|
return len(self.children) == 0
|
|
|
|
def ucb_score(self, total_visits, exploration_constant=math.sqrt(2)):
|
|
if self.visit_count == 0:
|
|
return float('inf')
|
|
avg_value = self.value_sum / self.visit_count
|
|
exploration_term = exploration_constant * self.prior * math.sqrt(total_visits) / (1 + self.visit_count)
|
|
return avg_value + exploration_term
|
|
|
|
class MCTS:
|
|
def __init__(self, prediction_network, dynamics_network, action_encoder, num_iterations=10, exploration_constant=math.sqrt(2)):
|
|
self.prediction_network = prediction_network
|
|
self.dynamics_network = dynamics_network
|
|
self.action_encoder = action_encoder
|
|
self.num_iterations = num_iterations
|
|
self.exploration_constant = exploration_constant
|
|
self.cache = {}
|
|
|
|
def search(self, root_state):
|
|
"""
|
|
Perform MCTS starting from the root state.
|
|
|
|
Args:
|
|
root_state (State): The root state from which to start the search.
|
|
|
|
Returns:
|
|
str: The best action to take from the root state.
|
|
"""
|
|
root_node = MCTSNode(state=root_state, thought_node=root_state.thought_node)
|
|
|
|
for _ in range(self.num_iterations):
|
|
node = self.select(root_node)
|
|
value = self.evaluate(node)
|
|
self.backpropagate(node, value)
|
|
|
|
best_action = self.best_action(root_node)
|
|
return best_action
|
|
|
|
def select(self, node):
|
|
while not node.is_leaf():
|
|
total_visits = sum(child.visit_count for child in node.children.values())
|
|
_, node = max(
|
|
node.children.items(),
|
|
key=lambda item: item[1].ucb_score(total_visits, self.exploration_constant)
|
|
)
|
|
return node
|
|
|
|
def evaluate(self, node):
|
|
|
|
state_representation = node.state.representation
|
|
policy_logits, value_estimate = self.prediction_network(state_representation)
|
|
value_estimate = value_estimate.item()
|
|
|
|
|
|
policy_probs = F.softmax(policy_logits, dim=-1).squeeze(0)
|
|
|
|
policy_probs = policy_probs[-1]
|
|
|
|
|
|
priors = {}
|
|
for child in node.thought_node.children:
|
|
action_name = child.name
|
|
action_idx = action_to_index.get(action_name, None)
|
|
if action_idx is not None and action_idx < policy_probs.size(0):
|
|
priors[action_name] = policy_probs[action_idx].item()
|
|
else:
|
|
priors[action_name] = 1.0 / len(node.thought_node.children)
|
|
|
|
|
|
node.expand(priors)
|
|
|
|
return value_estimate
|
|
|
|
def backpropagate(self, node, value):
|
|
while node is not None:
|
|
node.visit_count += 1
|
|
node.value_sum += value
|
|
node = node.parent
|
|
|
|
def best_action(self, root_node):
|
|
|
|
best_child = max(root_node.children.values(), key=lambda n: n.visit_count)
|
|
return best_child.action
|
|
|
|
class State:
|
|
def __init__(self, representation, dynamics_network, action_encoder, thought_node):
|
|
"""
|
|
Args:
|
|
representation (torch.Tensor): Encoded state representation, shape (batch_size, seq_len, state_dim)
|
|
dynamics_network (nn.Module): The Dynamics Network to predict next states
|
|
action_encoder (nn.Module): The Action Encoder to encode actions
|
|
thought_node (ThoughtNode): The current node in the Tree of Thought
|
|
"""
|
|
self.representation = representation
|
|
self.dynamics_network = dynamics_network
|
|
self.action_encoder = action_encoder
|
|
self.thought_node = thought_node
|
|
|
|
def apply_action(self, action):
|
|
"""
|
|
Apply an action to the current state to get a new state.
|
|
|
|
Args:
|
|
action (str): The action to apply (the name of the ThoughtNode)
|
|
|
|
Returns:
|
|
State: The new state after applying the action
|
|
"""
|
|
|
|
next_thought_node = None
|
|
for child in self.thought_node.children:
|
|
if child.name == action:
|
|
next_thought_node = child
|
|
break
|
|
if next_thought_node is None:
|
|
raise ValueError(f"Action '{action}' is not valid from the current thought node.")
|
|
|
|
|
|
action_index = torch.tensor([[action_to_index[action]]], device=self.representation.device)
|
|
action_embedding = self.action_encoder(action_index)
|
|
|
|
|
|
next_state_representation = self.dynamics_network(self.representation, action_embedding)
|
|
|
|
return State(
|
|
representation=next_state_representation,
|
|
dynamics_network=self.dynamics_network,
|
|
action_encoder=self.action_encoder,
|
|
thought_node=next_thought_node
|
|
)
|
|
|
|
class PPOAgent:
|
|
def __init__(self, policy_network, optimizer, clip_epsilon=0.2, entropy_coef=0.01, value_coef=0.5):
|
|
self.policy_network = policy_network
|
|
self.optimizer = optimizer
|
|
self.clip_epsilon = clip_epsilon
|
|
self.entropy_coef = entropy_coef
|
|
self.value_coef = value_coef
|
|
|
|
def compute_loss(self, states, old_log_probs, actions, returns, advantages):
|
|
|
|
policy_logits, value_estimates = self.policy_network(states)
|
|
batch_size, seq_len, num_actions = policy_logits.size()
|
|
|
|
|
|
policy_logits = policy_logits.reshape(-1, num_actions)
|
|
value_estimates = value_estimates.view(-1)
|
|
actions = actions.reshape(-1)
|
|
old_log_probs = old_log_probs.reshape(-1)
|
|
returns = returns.view(-1)
|
|
advantages = advantages.reshape(-1)
|
|
|
|
|
|
if value_estimates.size() != returns.size():
|
|
print(f"Shape mismatch: value_estimates shape: {value_estimates.size()}, returns shape: {returns.size()}")
|
|
value_estimates = value_estimates[:returns.size(0)]
|
|
|
|
|
|
new_log_probs_all = F.log_softmax(policy_logits, dim=-1)
|
|
new_log_probs = new_log_probs_all.gather(1, actions.unsqueeze(-1)).squeeze(-1)
|
|
|
|
|
|
ratios = torch.exp(new_log_probs - old_log_probs)
|
|
|
|
|
|
surr1 = ratios * advantages
|
|
surr2 = torch.clamp(ratios, 1 - self.clip_epsilon, 1 + self.clip_epsilon) * advantages
|
|
policy_loss = -torch.min(surr1, surr2).mean()
|
|
|
|
|
|
value_loss = F.mse_loss(value_estimates, returns)
|
|
|
|
|
|
entropy = -(new_log_probs * torch.exp(new_log_probs)).mean()
|
|
|
|
|
|
total_loss = policy_loss + self.value_coef * value_loss - self.entropy_coef * entropy
|
|
return total_loss
|
|
|
|
|
|
|
|
def infer(query, world_model_components, root_thought_node, tokenizer, max_length=20, inference_mode='world_model'):
|
|
"""
|
|
Perform inference given a query, utilizing the Tree of Thought and MCTS.
|
|
|
|
Args:
|
|
query (str): The input query or prompt.
|
|
world_model_components (tuple): Tuple containing the model components.
|
|
root_thought_node (ThoughtNode): The root node of the Tree of Thought.
|
|
tokenizer (transformers.PreTrainedTokenizer): The tokenizer used.
|
|
max_length (int): Maximum length for the generated sequence.
|
|
inference_mode (str): Inference mode ('world_model', 'without_world_model', 'world_model_tree_of_thought')
|
|
|
|
Returns:
|
|
List[str] or str: The sequence of actions (thoughts) selected or generated text.
|
|
"""
|
|
representation_network, dynamics_network, prediction_network, action_encoder, ppo_agent, model_transformer = world_model_components
|
|
|
|
|
|
input_ids = tokenizer.encode(query, return_tensors='pt').to(device)
|
|
attention_mask = (input_ids != tokenizer.pad_token_id).long()
|
|
|
|
if inference_mode == 'without_world_model':
|
|
|
|
with torch.no_grad():
|
|
generated_ids, entropies, variances = model_transformer.generate(src=input_ids, tokenizer=tokenizer, max_length=max_length, temperature=args.temperature)
|
|
generated_text = tokenizer.decode(generated_ids[0], skip_special_tokens=True)
|
|
return generated_text
|
|
|
|
else:
|
|
|
|
with torch.no_grad():
|
|
transformer_output = model_transformer(input_ids, input_ids)
|
|
|
|
initial_representation = representation_network(transformer_output)
|
|
initial_state = State(
|
|
representation=initial_representation,
|
|
dynamics_network=dynamics_network,
|
|
action_encoder=action_encoder,
|
|
thought_node=root_thought_node
|
|
)
|
|
|
|
if inference_mode == 'world_model_tree_of_thought':
|
|
|
|
mcts = MCTS(prediction_network, dynamics_network, action_encoder, num_iterations=args.mcts_iterations, exploration_constant=args.mcts_exploration_constant)
|
|
current_state = initial_state
|
|
thought_sequence = []
|
|
|
|
for _ in range(max_length):
|
|
best_action = mcts.search(current_state)
|
|
thought_sequence.append(best_action)
|
|
|
|
|
|
current_state = current_state.apply_action(best_action)
|
|
|
|
|
|
if len(current_state.thought_node.children) == 0:
|
|
break
|
|
|
|
return thought_sequence
|
|
else:
|
|
|
|
|
|
policy_logits, _ = prediction_network(initial_state.representation)
|
|
policy_probs = F.softmax(policy_logits, dim=-1)
|
|
|
|
top_actions = torch.argmax(policy_probs, dim=-1)
|
|
generated_actions = [index_to_action[idx.item()] for idx in top_actions[0]]
|
|
return generated_actions
|
|
|
|
def train_epoch_world_model(world_model_components, train_loader, optimizer, scheduler, scaler, args, model_transformer, state_dim, embed_dim, input_dim):
|
|
representation_network, dynamics_network, prediction_network, action_encoder, ppo_agent, _ = world_model_components
|
|
representation_network.train()
|
|
dynamics_network.train()
|
|
prediction_network.train()
|
|
action_encoder.train()
|
|
ppo_agent.policy_network.train()
|
|
|
|
total_loss = 0.0
|
|
optimizer.zero_grad()
|
|
print(f"Starting World Model training epoch with {len(train_loader)}batches...")
|
|
|
|
for i, batch in enumerate(train_loader):
|
|
print(f"Processing batch {i+1}/{len(train_loader)}...")
|
|
|
|
|
|
src_batch = batch['input_ids'].to(device)
|
|
tgt_batch = batch['labels'].to(device)
|
|
|
|
with torch.amp.autocast(device_type='cuda'):
|
|
print("Forward pass through Transformer (frozen)...")
|
|
with torch.no_grad():
|
|
transformer_output = model_transformer(src_batch, tgt_batch[:, :-1])
|
|
|
|
|
|
state_representation = representation_network(transformer_output)
|
|
|
|
|
|
true_actions = tgt_batch[:, :-1]
|
|
action_sequences = true_actions
|
|
|
|
|
|
action_embeddings = action_encoder(action_sequences)
|
|
|
|
|
|
predicted_next_state_batch = dynamics_network(state_representation, action_embeddings)
|
|
|
|
|
|
policy_logits, value_estimates = prediction_network(predicted_next_state_batch)
|
|
|
|
|
|
|
|
true_policy = F.one_hot(true_actions, num_classes=input_dim).float()
|
|
true_value = torch.zeros_like(value_estimates).to(device)
|
|
|
|
|
|
actions_selected = true_actions
|
|
old_log_probs = torch.zeros_like(actions_selected, dtype=torch.float32).to(device)
|
|
returns = torch.zeros_like(actions_selected, dtype=torch.float32).to(device)
|
|
advantages = torch.zeros_like(actions_selected, dtype=torch.float32).to(device)
|
|
|
|
|
|
ppo_loss = ppo_agent.compute_loss(state_representation, old_log_probs, actions_selected, returns, advantages)
|
|
|
|
|
|
z_i = state_representation.view(-1, state_dim)
|
|
z_j = F.dropout(z_i, p=0.1, training=True)
|
|
info_nce = InfoNCE_Loss()(z_i, z_j)
|
|
|
|
|
|
covariance = CovarianceRegularization()(predicted_next_state_batch.view(-1, predicted_next_state_batch.size(-1)))
|
|
dynamics_loss = DynamicsPerformanceLoss()(state_representation, predicted_next_state_batch)
|
|
perturbed_next_state = predicted_next_state_batch + torch.randn_like(predicted_next_state_batch) * 0.01
|
|
thought_loss = ThoughtConsistencyLoss()(predicted_next_state_batch, perturbed_next_state)
|
|
pv_loss = PolicyValueJointLoss()(policy_logits, true_policy, value_estimates.squeeze(-1), true_value.squeeze(-1))
|
|
action_diversity = ActionDiversityReward()(action_embeddings.view(-1, embed_dim))
|
|
mcts_best_values = torch.zeros(actions_selected.size(0)).to(device)
|
|
etv = ExpectedThoughtValueLoss()(mcts_best_values)
|
|
visit_counts = torch.ones(actions_selected.size(0), policy_logits.size(-1)).to(device)
|
|
exploration = ExplorationRegularization()(visit_counts)
|
|
old_policy = F.softmax(policy_logits.detach(), dim=-1)
|
|
new_policy = F.softmax(policy_logits, dim=-1)
|
|
kl_loss = KL_DivergenceLoss()(old_policy, new_policy)
|
|
|
|
|
|
loss = (
|
|
ppo_loss +
|
|
info_nce +
|
|
covariance +
|
|
dynamics_loss +
|
|
thought_loss +
|
|
pv_loss +
|
|
action_diversity +
|
|
etv +
|
|
exploration +
|
|
kl_loss
|
|
)
|
|
loss = loss / args.accumulation_steps
|
|
|
|
print("Backward pass...")
|
|
scaler.scale(loss).backward()
|
|
|
|
if (i + 1) % args.accumulation_steps == 0 or (i + 1) == len(train_loader):
|
|
print("Gradient clipping...")
|
|
scaler.unscale_(optimizer)
|
|
torch.nn.utils.clip_grad_norm_(
|
|
[param for group in optimizer.param_groups for param in group['params']],
|
|
args.max_grad_norm
|
|
)
|
|
|
|
print("Optimizer step...")
|
|
scaler.step(optimizer)
|
|
scaler.update()
|
|
|
|
print("Zeroing gradients...")
|
|
optimizer.zero_grad()
|
|
|
|
print("Updating learning rate...")
|
|
scheduler.step()
|
|
|
|
total_loss += loss.item() * args.accumulation_steps
|
|
print(f"Batch {i+1} completed. Current loss: {loss.item():.4f}")
|
|
|
|
avg_loss = total_loss / len(train_loader)
|
|
print(f"World Model training epoch completed. Average loss: {avg_loss:.4f}")
|
|
return avg_loss
|
|
|
|
|
|
def train_epoch_language_model(model, train_loader, optimizer, scheduler, scaler, args):
|
|
model.train()
|
|
total_loss = 0.0
|
|
optimizer.zero_grad()
|
|
print(f"Starting Language Model training epoch with {len(train_loader)} batches...")
|
|
|
|
for i, batch in enumerate(train_loader):
|
|
input_ids = batch['input_ids'].to(device)
|
|
labels = batch['labels'].to(device)
|
|
|
|
with autocast():
|
|
outputs = model(input_ids, input_ids)
|
|
logits = outputs.view(-1, outputs.size(-1))
|
|
labels = labels.view(-1)
|
|
loss = F.cross_entropy(logits, labels, ignore_index=model.embedding.padding_idx)
|
|
loss = loss / args.accumulation_steps
|
|
|
|
scaler.scale(loss).backward()
|
|
|
|
if (i + 1) % args.accumulation_steps == 0 or (i + 1) == len(train_loader):
|
|
scaler.unscale_(optimizer)
|
|
torch.nn.utils.clip_grad_norm_(
|
|
[param for group in optimizer.param_groups for param in group['params']],
|
|
args.max_grad_norm
|
|
)
|
|
scaler.step(optimizer)
|
|
scaler.update()
|
|
optimizer.zero_grad()
|
|
scheduler.step()
|
|
|
|
total_loss += loss.item() * args.accumulation_steps
|
|
print(f"Batch {i + 1} completed. Current loss: {loss.item():.4f}")
|
|
|
|
avg_loss = total_loss / len(train_loader)
|
|
print(f"Language Model training epoch completed. Average loss: {avg_loss:.4f}")
|
|
return avg_loss
|
|
|
|
|
|
|
|
def main():
|
|
args = parse_args()
|
|
print("Arguments parsed successfully.")
|
|
|
|
|
|
os.makedirs(args.save_dir, exist_ok=True)
|
|
print(f"Save directory created: {args.save_dir}")
|
|
|
|
|
|
print("Loading tokenizer...")
|
|
tokenizer = AutoTokenizer.from_pretrained(args.model_name)
|
|
if tokenizer.pad_token is None:
|
|
tokenizer.pad_token = tokenizer.eos_token
|
|
print("Tokenizer loaded successfully.")
|
|
|
|
|
|
padding_idx = tokenizer.pad_token_id
|
|
input_dim = len(tokenizer)
|
|
|
|
|
|
print("Initializing Transformer model...")
|
|
model_transformer = Transformer(
|
|
input_dim=input_dim,
|
|
d_model=128,
|
|
num_heads=4,
|
|
num_layers=4,
|
|
d_ff=256,
|
|
num_experts=2,
|
|
output_dim=input_dim,
|
|
dropout=0.1,
|
|
top_k=2
|
|
).to(device)
|
|
model_transformer.train()
|
|
print("Transformer model initialized on device.")
|
|
|
|
|
|
d_model = 128
|
|
state_dim = 128
|
|
action_dim = d_model
|
|
hidden_dim = 256
|
|
vocab_dim = input_dim
|
|
embed_dim = d_model
|
|
|
|
|
|
representation_network = RepresentationNetwork(vocab_dim, d_model, state_dim).to(device)
|
|
dynamics_network = DynamicsNetwork(state_dim, action_dim, hidden_dim).to(device)
|
|
prediction_network = PredictionNetwork(state_dim, input_dim, 1).to(device)
|
|
action_encoder = ActionEncoder(input_dim, action_dim).to(device)
|
|
|
|
|
|
ppo_agent = PPOAgent(
|
|
policy_network=prediction_network,
|
|
optimizer=optim.AdamW(prediction_network.parameters(), lr=args.learning_rate),
|
|
clip_epsilon=0.2,
|
|
entropy_coef=0.01,
|
|
value_coef=0.5
|
|
)
|
|
|
|
|
|
world_model_components = (representation_network, dynamics_network, prediction_network, action_encoder, ppo_agent, model_transformer)
|
|
|
|
if args.mode == 'train':
|
|
print("Loading and preprocessing data...")
|
|
train_loader, eval_loader = load_data(args, tokenizer)
|
|
print("Data loaded and preprocessed successfully.")
|
|
|
|
|
|
optimizer = optim.AdamW(
|
|
list(representation_network.parameters()) +
|
|
list(dynamics_network.parameters()) +
|
|
list(prediction_network.parameters()) +
|
|
list(action_encoder.parameters()),
|
|
lr=args.learning_rate, weight_decay=args.weight_decay
|
|
) if args.train_mode == 'world_model' else optim.AdamW(model_transformer.parameters(), lr=args.learning_rate)
|
|
scheduler = CosineAnnealingLR(optimizer, T_max=args.num_epochs)
|
|
scaler = GradScaler()
|
|
|
|
print(f"Starting {args.train_mode} training...")
|
|
|
|
for epoch in range(args.num_epochs):
|
|
if args.train_mode == 'world_model':
|
|
avg_loss = train_epoch_world_model(
|
|
world_model_components,
|
|
train_loader,
|
|
optimizer,
|
|
scheduler,
|
|
scaler,
|
|
args,
|
|
model_transformer,
|
|
state_dim,
|
|
embed_dim,
|
|
input_dim
|
|
)
|
|
else:
|
|
avg_loss = train_epoch_language_model(
|
|
model_transformer,
|
|
train_loader,
|
|
optimizer,
|
|
scheduler,
|
|
scaler,
|
|
args
|
|
)
|
|
|
|
print(f"{args.train_mode.capitalize()} training epoch {epoch + 1} completed. Average loss: {avg_loss:.4f}")
|
|
|
|
if args.train_mode == 'world_model':
|
|
save_all_models(model_transformer, representation_network, dynamics_network, prediction_network, action_encoder, args.save_dir, epoch + 1)
|
|
print(f"Models saved for epoch {epoch + 1}")
|
|
else:
|
|
torch.save(model_transformer.state_dict(), os.path.join(args.save_dir, f'language_model_epoch_{epoch + 1}.pt'))
|
|
print(f"Language model saved for epoch {epoch + 1}")
|
|
|
|
print("Training completed.")
|
|
|
|
elif args.mode == 'inference':
|
|
|
|
tree_root = build_tree_of_thought()
|
|
|
|
action_list = []
|
|
traverse_tree(tree_root, action_list)
|
|
|
|
|
|
global action_to_index, index_to_action
|
|
action_to_index = {action: idx for idx, action in enumerate(action_list)}
|
|
index_to_action = {idx: action for action, idx in action_to_index.items()}
|
|
action_vocab_size = len(action_list)
|
|
|
|
|
|
action_encoder = ActionEncoder(action_vocab_size, action_dim).to(device)
|
|
prediction_network = PredictionNetwork(state_dim, action_vocab_size, 1).to(device)
|
|
|
|
|
|
|
|
|
|
model_transformer.load_state_dict(torch.load(os.path.join(args.save_dir, 'transformer_model_epoch_2.pt')))
|
|
representation_network.load_state_dict(torch.load(os.path.join(args.save_dir, 'representation_network_epoch_2.pt')))
|
|
dynamics_network.load_state_dict(torch.load(os.path.join(args.save_dir, 'dynamics_network_epoch_2.pt')))
|
|
saved_state_dict = torch.load(os.path.join(args.save_dir, 'prediction_network_epoch_2.pt'))
|
|
prediction_network.policy_head = nn.Linear(prediction_network.state_dim, 50257)
|
|
prediction_network.load_state_dict(saved_state_dict, strict=False)
|
|
|
|
|
|
prediction_network.policy_head = nn.Linear(prediction_network.state_dim, 158).to(device)
|
|
|
|
|
|
|
|
action_encoder.load_state_dict(torch.load(os.path.join(args.save_dir, 'action_encoder_epoch_2.pt')))
|
|
|
|
|
|
world_model_components = (representation_network, dynamics_network, prediction_network, action_encoder, ppo_agent, model_transformer)
|
|
|
|
|
|
if not args.query:
|
|
args.query = input("Please enter your query: ")
|
|
|
|
result = infer(args.query, world_model_components, tree_root, tokenizer, inference_mode=args.inference_mode)
|
|
|
|
if args.inference_mode == 'without_world_model':
|
|
print("Generated Text:")
|
|
print(result)
|
|
else:
|
|
print("Generated Thought Sequence:")
|
|
for thought in result:
|
|
print(thought)
|
|
|
|
if __name__ == '__main__':
|
|
main()
|
|
|