add a demo (scifi-train)

llama-3/README.md

@@ -1,5 +1,7 @@
 # 中文微调llama-3
 
+这个目录是用来存放中文微调llama-3的代码的。与其他目录无关。
+
 目前最省GPU最快的方法微调llama-3就是通过usloth的方法，这个方法是在llama-3的基础上，他们预先量化到了4bit，减少微调时所需的内存。
 
 这个方法的优点是，不需要重新训练模型，只需要下载预训练模型，然后微调即可。

scifi-demo/README.md

@@ -0,0 +1,13 @@
+## About this repo
+
+这个目录是我在B站、抖音上发布视频里的一个demo。
+
+通过我们跟目录下一样的手写的模型来实现一个科幻小说微调原理。教学作用，代码不适用于生产环境。
+
+### 准备数据集
+
+准备数据集的部分需要预处理，并存放在 `/data` 子文件夹下。
+
+ - 训练数据集 `tools_download`工具代码中的第18～35行。先把zip包解压，然后合并所有的小说txt文件到单一的txt文件当中。
+
+ - 微调数据集 `tools_download`工具代码中的第6～14行。从huggingface上下载一个带微调指令的数据集。

scifi-demo/finetune-scifi.py

@@ -0,0 +1,101 @@
+"""
+Fine-tune a model
+"""
+import os
+import sys
+import pickle
+from contextlib import nullcontext
+import torch
+import tiktoken
+from aim import Run
+from model import Model
+import json
+
+
+# Hyperparameters
+batch_size = 8  # How many batches per training step
+context_length = 128  # Length of the token chunk each batch
+max_iters = 1000  # Total of training iterations <- Change this to smaller number for testing
+learning_rate = 1e-4  # 0.001
+eval_interval = 10  # How often to evaluate
+eval_iters = 10  # Number of iterations to average for evaluation
+device = 'cuda' if torch.cuda.is_available() else 'cpu'  # Use GPU if it's available.
+TORCH_SEED = 1337
+torch.manual_seed(TORCH_SEED)
+
+
+# 准备训练数据
+with open('data/scifi-finetune.json', 'r') as file:
+    alpaca = json.load(file)
+    text = alpaca[1000:5001]
+
+# print(text)
+# sys.exit(0)
+
+# Using TikToken (Same as GPT3) to tokenize the source text
+encoding = tiktoken.get_encoding("cl100k_base")
+tokenized_text = encoding.encode(str(text))
+tokenized_text = torch.tensor(tokenized_text, dtype=torch.long, device=device)  # 将77,919个tokens 转换到Pytorch张量中
+
+total_tokens = encoding.encode_ordinary(str(text))
+print(f"数据集合计有 {len(total_tokens):,} tokens")
+
+
+# Split train and validation
+train_size = int(len(tokenized_text) * 0.9)
+train_data = tokenized_text[:train_size]
+val_data = tokenized_text[train_size:]
+
+
+# Initialize the model
+model = Model()
+model.load_state_dict(torch.load('model/model-scifi.pt'))
+model.to(device)
+
+# get batch
+def get_batch(split: str):
+    data = train_data if split == 'train' else val_data
+    idxs = torch.randint(low=0, high=len(data) - context_length, size=(batch_size,))
+    x = torch.stack([data[idx:idx + context_length] for idx in idxs]).to(device)
+    y = torch.stack([data[idx + 1:idx + context_length + 1] for idx in idxs]).to(device)
+    return x, y
+
+
+# calculate the loss
+@torch.no_grad()
+def estimate_loss():
+    out = {}
+    model.eval()
+    for split in ['train', 'valid']:
+        losses = torch.zeros(eval_iters)
+        for k in range(eval_iters):
+            x_batch, y_batch = get_batch(split)
+            logits, loss = model(x_batch, y_batch)
+            losses[k] = loss.item()
+        out[split] = losses.mean()
+    model.train()
+    return out
+
+
+# Create the optimizer
+optimizer = torch.optim.AdamW(model.parameters(), lr=learning_rate)
+tracked_losses = list()
+for step in range(max_iters):
+    if step % eval_iters == 0 or step == max_iters - 1:
+        losses = estimate_loss()
+        tracked_losses.append(losses)
+        print('Step:', step, 'Training Loss:', round(losses['train'].item(), 3), 'Validation Loss:', round(losses['valid'].item(), 3))
+
+    xb, yb = get_batch('train')
+    logits, loss = model(xb, yb)
+    optimizer.zero_grad(set_to_none=True)
+    loss.backward()
+    optimizer.step()
+
+# Save the model
+torch.save(model.state_dict(), 'model/model-scifi-finetune.pt')
+
+
+
+
+

scifi-demo/inference-scifi.py

@@ -0,0 +1,42 @@
+# -*- coding: utf-8 -*-
+"""
+Sample from a trained model
+"""
+import os
+import pickle
+from contextlib import nullcontext
+import torch
+import tiktoken
+from model import Model
+
+
+# Hyperparameters
+device = 'cuda' if torch.cuda.is_available() else 'cpu'
+TORCH_SEED = 1337
+torch.manual_seed(TORCH_SEED)
+torch.cuda.manual_seed(TORCH_SEED)
+
+
+encoding = tiktoken.get_encoding("cl100k_base")
+
+
+# Initiate from trained model
+model = Model()
+model.load_state_dict(torch.load('model/model-scifi.pt'))
+model.eval()
+model.to(device)
+
+# start = 'Write a short story about Sam Altman.'
+start = 'Sam Altman was born in'
+start_ids = encoding.encode(start)
+x = (torch.tensor(start_ids, dtype=torch.long, device=device)[None, ...])
+
+# run generation
+with torch.no_grad():
+    y = model.generate(x, max_new_tokens=500)
+    print('---------------')
+    print(encoding.decode(y[0].tolist()))
+    print('---------------')
+
+
+

scifi-demo/model.py

@@ -0,0 +1,138 @@
+import math
+import torch
+import torch.nn as nn
+from torch.nn import functional as F
+import tiktoken
+
+# Hyperparameters
+context_length = 128  # Length of the token chunk each batch
+d_model = 512  # The size of our model token embeddings
+num_blocks = 12  # Number of transformer blocks
+num_heads = 8  # Number of heads in Multi-head attention
+dropout = 0.1  # Dropout rate
+device = 'cuda' if torch.cuda.is_available() else 'cpu'  # Use GPU if it's available.
+TORCH_SEED = 1337
+torch.manual_seed(TORCH_SEED)
+
+# Define feed forward network
+class FeedForwardNetwork(nn.Module):
+    def __init__(self):
+        super().__init__()
+        self.ffn = nn.Sequential(
+            nn.Linear(d_model, d_model * 4),
+            nn.ReLU(),
+            nn.Linear(d_model * 4, d_model),
+            nn.Dropout(dropout)
+        )
+
+    def forward(self, x):
+        return self.ffn(x)
+
+
+# Define Scaled Dot Product Attention
+class Attention(nn.Module):
+    def __init__(self):
+        super().__init__()
+        self.Wq = nn.Linear(d_model, d_model // num_heads, bias=False)
+        self.Wk = nn.Linear(d_model, d_model // num_heads, bias=False)
+        self.Wv = nn.Linear(d_model, d_model // num_heads, bias=False)
+        self.register_buffer('mask', torch.tril(torch.ones(context_length, context_length)))
+        self.dropout = nn.Dropout(dropout)
+
+    def forward(self, x):
+        B, T, C = x.shape
+        q = self.Wq(x)
+        k = self.Wk(x)
+        v = self.Wv(x)
+
+        weights = (q @ k.transpose(-2, -1)) / math.sqrt(d_model // num_heads)
+        weights = weights.masked_fill(self.mask[:T, :T] == 0, float('-inf'))
+        weights = F.softmax(weights, dim=-1)
+        weights = self.dropout(weights)
+
+        output = weights @ v
+
+        return output
+
+
+# Define Multi-head Attention
+class MultiHeadAttention(nn.Module):
+    def __init__(self):
+        super().__init__()
+        self.heads = nn.ModuleList([Attention() for _ in range(num_heads)])
+        self.projection_layer = nn.Linear(d_model, d_model)
+        self.dropout = nn.Dropout(dropout)
+
+    def forward(self, x):
+        head_outputs = [head(x) for head in self.heads]
+        head_outputs = torch.cat(head_outputs, dim=-1)
+        out = self.dropout(self.projection_layer(head_outputs))
+        return out
+
+
+# Define Transformer Block
+class TransformerBlock(nn.Module):
+    def __init__(self):
+        super().__init__()
+        self.ln1 = nn.LayerNorm(d_model)
+        self.ln2 = nn.LayerNorm(d_model)
+        self.mha = MultiHeadAttention()
+        self.ffn = FeedForwardNetwork()
+
+    def forward(self, x):
+        x = x + self.mha(self.ln1(x))
+        x = x + self.ffn(self.ln2(x))
+        return x
+
+
+# Define the model
+class Model(nn.Module):
+    def __init__(self, max_token_value=100256): # if not passed, force to be default tiktoken cl100k vocab size
+        super().__init__()
+        self.token_embedding_lookup_table = nn.Embedding(max_token_value, d_model)
+        self.transformer_blocks = nn.Sequential(*(
+                [TransformerBlock() for _ in range(num_blocks)] +
+                [nn.LayerNorm(d_model)]
+        ))
+        self.model_out_linear_layer = nn.Linear(d_model, max_token_value)
+
+    def forward(self, idx, targets=None):
+        B, T = idx.shape
+        position_encoding_lookup_table = torch.zeros(context_length, d_model, device=device)
+        position = torch.arange(0, context_length, dtype=torch.float).unsqueeze(1)
+        div_term = torch.exp(torch.arange(0, d_model, 2).float() * (-math.log(10000.0) / d_model))
+        position_encoding_lookup_table[:, 0::2] = torch.sin(position * div_term)
+        position_encoding_lookup_table[:, 1::2] = torch.cos(position * div_term)
+        # change position_encoding_lookup_table from (context_length, d_model) to (T, d_model)
+        position_embedding = position_encoding_lookup_table[:T, :].to(device)
+        x = self.token_embedding_lookup_table(idx) + position_embedding
+        x = self.transformer_blocks(x)
+        # get the final logits
+        logits = self.model_out_linear_layer(x)
+
+        if targets is not None:
+            B, T, C = logits.shape
+            logits_reshaped = logits.view(B * T, C)
+            targets_reshaped = targets.view(B * T)
+            loss = F.cross_entropy(input=logits_reshaped, target=targets_reshaped)
+        else:
+            loss = None
+        return logits, loss
+
+    def generate(self, idx, max_new_tokens=100):
+        # idx is (B,T) array of indices in the current context
+        for _ in range(max_new_tokens):
+            # Crop idx to the max size of our positional embeddings table
+            idx_crop = idx[:, -context_length:]
+            # Get predictions
+            logits, loss = self.forward(idx_crop)
+            # Get the last time step from logits where the dimensions of the logits are (B,T,C)
+            logits_last_timestep = logits[:, -1, :]
+            # Apply softmax to get probabilities
+            probs = F.softmax(input=logits_last_timestep, dim=-1)
+            # Sample from the probabilities' distribution.
+            idx_next = torch.multinomial(input=probs, num_samples=1)
+            # Append the sampled indexes idx_next to idx
+            idx = torch.cat((idx, idx_next), dim=1)
+        return idx
+

scifi-demo/show-parameters.py

@@ -0,0 +1,11 @@
+import torch
+from model import Model
+
+model = Model()
+state_dict = torch.load('model/model-scifi.pt')
+model.load_state_dict(state_dict)
+
+# Calculate the number of parameters
+total_params = sum(p.numel() for p in model.parameters() if p.requires_grad)
+
+print(f"模型参数量为： {total_params:,}")

scifi-demo/tools_download.py

@@ -0,0 +1,35 @@
+import os
+import sys
+import requests
+import glob
+
+def download_file(url, save_path):
+    response = requests.get(url, stream=True)
+    with open(save_path, 'wb') as file:
+        for chunk in response.iter_content(chunk_size=8192):
+            file.write(chunk)
+
+url = "https://huggingface.co/datasets/zxbsmk/webnovel_cn/resolve/main/novel_cn_token512_50k.json?download=true"
+save_path = "data/scifi-finetune.json"
+download_file(url, save_path)
+
+sys.exit(0)
+
+# 小说模型训练数据集下载链接: https://pan.baidu.com/s/1bC8fH8hyt28L9pV3fjOHIQ 提取码: 9i9g
+def find_txt_files(directory):
+    return glob.glob(os.path.join(directory, '**', '*.txt'), recursive=True)
+
+def concatenate_txt_files(files, output_file):
+    with open(output_file, 'w') as outfile:
+        for file in files:
+            with open(file, 'r') as infile:
+                outfile.write(infile.read() + '\n')  # Adds a newline between files
+
+directory = 'data'
+output_file = 'data/scifi.txt'
+
+# Find all .txt files
+txt_files = find_txt_files(directory)
+
+# Concatenate all found .txt files into one
+concatenate_txt_files(txt_files, output_file)