[booster] init module structure and definition (hpcaitech#3056)

applications/ChatGPT/chatgpt/models/llama/llama.py

@@ -0,0 +1,316 @@
+# Copyright (c) Meta Platforms, Inc. and affiliates.
+# This software may be used and distributed according to the terms of the GNU General Public License version 3.
+
+import json
+import math
+import time
+from dataclasses import dataclass
+from pathlib import Path
+from typing import List, Optional, Tuple
+
+import torch
+import torch.nn.functional as F
+from torch import nn
+
+from colossalai.context.parallel_mode import ParallelMode
+from colossalai.core import global_context as gpc
+from colossalai.nn.layer import Embedding1D, Linear1D_Col, Linear1D_Row
+
+from .llama_tokenizer import LLaMATokenizer
+
+
+@dataclass
+class LLaMAConfig:
+    dim: int = 512
+    n_layers: int = 8
+    n_heads: int = 8
+    vocab_size: int = -1    # defined later by tokenizer
+    multiple_of: int = 256    # make SwiGLU hidden layer size multiple of large power of 2
+    norm_eps: float = 1e-5
+
+    max_batch_size: int = 32
+    max_seq_len: int = 1024
+
+
+class RMSNorm(nn.Module):
+
+    def __init__(self, dim: int, eps: float = 1e-6):
+        super().__init__()
+        self.eps = eps
+        self.weight = nn.Parameter(torch.ones(dim))
+
+    def _norm(self, x):
+        return x * torch.rsqrt(x.pow(2).mean(-1, keepdim=True) + self.eps)
+
+    def forward(self, x):
+        output = self._norm(x.float()).type_as(x)
+        return output * self.weight
+
+
+def precompute_freqs_cis(dim: int, end: int, theta: float = 10000.0):
+    freqs = 1.0 / (theta**(torch.arange(0, dim, 2)[:(dim // 2)].float() / dim))
+    t = torch.arange(end, device=freqs.device)    # type: ignore
+    freqs = torch.outer(t, freqs).float()    # type: ignore
+    freqs_cis = torch.polar(torch.ones_like(freqs), freqs)    # complex64
+    return freqs_cis
+
+
+def reshape_for_broadcast(freqs_cis: torch.Tensor, x: torch.Tensor):
+    ndim = x.ndim
+    assert 0 <= 1 < ndim
+    assert freqs_cis.shape == (x.shape[1], x.shape[-1])
+    shape = [d if i == 1 or i == ndim - 1 else 1 for i, d in enumerate(x.shape)]
+    return freqs_cis.view(*shape)
+
+
+def apply_rotary_emb(
+    xq: torch.Tensor,
+    xk: torch.Tensor,
+    freqs_cis: torch.Tensor,
+) -> Tuple[torch.Tensor, torch.Tensor]:
+    xq_ = torch.view_as_complex(xq.float().reshape(*xq.shape[:-1], -1, 2))
+    xk_ = torch.view_as_complex(xk.float().reshape(*xk.shape[:-1], -1, 2))
+    freqs_cis = reshape_for_broadcast(freqs_cis, xq_)
+    xq_out = torch.view_as_real(xq_ * freqs_cis).flatten(3)
+    xk_out = torch.view_as_real(xk_ * freqs_cis).flatten(3)
+    return xq_out.type_as(xq), xk_out.type_as(xk)
+
+
+class Attention(nn.Module):
+
+    def __init__(self, args: LLaMAConfig):
+        super().__init__()
+
+        self.n_local_heads = args.n_heads // gpc.get_world_size(ParallelMode.PARALLEL_1D)
+        self.head_dim = args.dim // args.n_heads
+
+        self.wq = Linear1D_Col(
+            args.dim,
+            args.n_heads * self.head_dim,
+            bias=False,
+            gather_output=False,
+            init_method=lambda x: x,
+        )
+        self.wk = Linear1D_Col(
+            args.dim,
+            args.n_heads * self.head_dim,
+            bias=False,
+            gather_output=False,
+            init_method=lambda x: x,
+        )
+        self.wv = Linear1D_Col(
+            args.dim,
+            args.n_heads * self.head_dim,
+            bias=False,
+            gather_output=False,
+            init_method=lambda x: x,
+        )
+        self.wo = Linear1D_Row(
+            args.n_heads * self.head_dim,
+            args.dim,
+            bias=False,
+            input_is_parallel=True,
+            init_method=lambda x: x,
+        )
+
+        self.cache_k = torch.zeros((args.max_batch_size, args.max_seq_len, self.n_local_heads, self.head_dim)).cuda()
+        self.cache_v = torch.zeros((args.max_batch_size, args.max_seq_len, self.n_local_heads, self.head_dim)).cuda()
+
+    def forward(self, x: torch.Tensor, start_pos: int, freqs_cis: torch.Tensor, mask: Optional[torch.Tensor]):
+        bsz, seqlen, _ = x.shape
+        xq, xk, xv = self.wq(x), self.wk(x), self.wv(x)
+
+        xq = xq.view(bsz, seqlen, self.n_local_heads, self.head_dim)
+        xk = xk.view(bsz, seqlen, self.n_local_heads, self.head_dim)
+        xv = xv.view(bsz, seqlen, self.n_local_heads, self.head_dim)
+
+        xq, xk = apply_rotary_emb(xq, xk, freqs_cis=freqs_cis)
+
+        self.cache_k = self.cache_k.to(xq)
+        self.cache_v = self.cache_v.to(xq)
+
+        self.cache_k[:bsz, start_pos:start_pos + seqlen] = xk
+        self.cache_v[:bsz, start_pos:start_pos + seqlen] = xv
+
+        keys = self.cache_k[:bsz, :start_pos + seqlen]
+        values = self.cache_v[:bsz, :start_pos + seqlen]
+
+        xq = xq.transpose(1, 2)
+        keys = keys.transpose(1, 2)
+        values = values.transpose(1, 2)
+        scores = torch.matmul(xq, keys.transpose(2, 3)) / math.sqrt(self.head_dim)
+        if mask is not None:
+            scores = scores + mask    # (bs, n_local_heads, slen, cache_len + slen)
+        scores = F.softmax(scores.float(), dim=-1).type_as(xq)
+        output = torch.matmul(scores, values)    # (bs, n_local_heads, slen, head_dim)
+        output = output.transpose(1, 2).contiguous().view(bsz, seqlen, -1)
+
+        return self.wo(output)
+
+
+class FeedForward(nn.Module):
+
+    def __init__(
+        self,
+        dim: int,
+        hidden_dim: int,
+        multiple_of: int,
+    ):
+        super().__init__()
+        hidden_dim = int(2 * hidden_dim / 3)
+        hidden_dim = multiple_of * ((hidden_dim + multiple_of - 1) // multiple_of)
+
+        self.w1 = Linear1D_Col(dim, hidden_dim, bias=False, gather_output=False, init_method=lambda x: x)
+        self.w2 = Linear1D_Row(hidden_dim, dim, bias=False, input_is_parallel=True, init_method=lambda x: x)
+        self.w3 = Linear1D_Col(dim, hidden_dim, bias=False, gather_output=False, init_method=lambda x: x)
+
+    def forward(self, x):
+        return self.w2(F.silu(self.w1(x)) * self.w3(x))
+
+
+class TransformerBlock(nn.Module):
+
+    def __init__(self, layer_id: int, args: LLaMAConfig):
+        super().__init__()
+        self.n_heads = args.n_heads
+        self.dim = args.dim
+        self.head_dim = args.dim // args.n_heads
+        self.attention = Attention(args)
+        self.feed_forward = FeedForward(dim=args.dim, hidden_dim=4 * args.dim, multiple_of=args.multiple_of)
+        self.layer_id = layer_id
+        self.attention_norm = RMSNorm(args.dim, eps=args.norm_eps)
+        self.ffn_norm = RMSNorm(args.dim, eps=args.norm_eps)
+
+    def forward(self, x: torch.Tensor, start_pos: int, freqs_cis: torch.Tensor, mask: Optional[torch.Tensor]):
+        h = x + self.attention.forward(self.attention_norm(x), start_pos, freqs_cis, mask)
+        out = h + self.feed_forward.forward(self.ffn_norm(h))
+        return out
+
+
+class LLaMA(nn.Module):
+
+    def __init__(self, config: LLaMAConfig):
+        super().__init__()
+        self.config = config
+        self.vocab_size = config.vocab_size
+        self.n_layers = config.n_layers
+
+        self.tok_embeddings = Embedding1D(config.vocab_size, config.dim, init_method=lambda x: x)
+
+        self.layers = torch.nn.ModuleList()
+        for layer_id in range(config.n_layers):
+            self.layers.append(TransformerBlock(layer_id, config))
+
+        self.norm = RMSNorm(config.dim, eps=config.norm_eps)
+        self.output = Linear1D_Col(config.dim, config.vocab_size, bias=False, init_method=lambda x: x)
+
+        self.freqs_cis = precompute_freqs_cis(self.config.dim // self.config.n_heads, self.config.max_seq_len * 2)
+
+    # @torch.inference_mode()
+    def forward(self, tokens: torch.Tensor, start_pos: int):
+        _bsz, seqlen = tokens.shape
+        h = self.tok_embeddings(tokens)
+        self.freqs_cis = self.freqs_cis.to(h.device)
+        freqs_cis = self.freqs_cis[start_pos:start_pos + seqlen]
+
+        mask = None
+        if seqlen > 1:
+            mask = torch.full((1, 1, seqlen, seqlen), float("-inf"), device=tokens.device)
+            mask = torch.triu(mask, diagonal=start_pos + 1).type_as(h)
+
+        for layer in self.layers:
+            h = layer(h, start_pos, freqs_cis, mask)
+        h = self.norm(h)
+        output = self.output(h[:, -1, :])    # only compute last logits
+        return output.float()
+
+
+class LLaMAGenerator:
+
+    def __init__(self, model: LLaMA, tokenizer: LLaMATokenizer):
+        self.model = model
+        self.tokenizer = tokenizer
+
+    def generate(
+        self,
+        prompts: List[str],
+        max_gen_len: int,
+        temperature: float = 0.8,
+        top_p: float = 0.95,
+    ) -> List[str]:
+        bsz = len(prompts)
+        config = self.model.config
+        assert bsz <= config.max_batch_size, (bsz, config.max_batch_size)
+
+        prompt_tokens = [self.tokenizer.encode(x, bos=True, eos=False) for x in prompts]
+
+        min_prompt_size = min([len(t) for t in prompt_tokens])
+        max_prompt_size = max([len(t) for t in prompt_tokens])
+
+        total_len = min(config.max_seq_len, max_gen_len + max_prompt_size)
+
+        tokens = torch.full((bsz, total_len), self.tokenizer.pad_id).cuda().long()
+        for k, t in enumerate(prompt_tokens):
+            tokens[k, :len(t)] = torch.tensor(t).long()
+        input_text_mask = tokens != self.tokenizer.pad_id
+        start_pos = min_prompt_size
+        prev_pos = 0
+        for cur_pos in range(start_pos, total_len):
+            logits = self.model.forward(tokens[:, prev_pos:cur_pos], prev_pos)
+            if temperature > 0:
+                probs = torch.softmax(logits / temperature, dim=-1)
+                next_token = sample_top_p(probs, top_p)
+            else:
+                next_token = torch.argmax(logits, dim=-1)
+            next_token = next_token.reshape(-1)
+            # only replace token if prompt has already been generated
+            next_token = torch.where(input_text_mask[:, cur_pos], tokens[:, cur_pos], next_token)
+            tokens[:, cur_pos] = next_token
+            prev_pos = cur_pos
+
+        decoded = []
+        for i, t in enumerate(tokens.tolist()):
+            # cut to max gen len
+            t = t[:len(prompt_tokens[i]) + max_gen_len]
+            # cut to eos tok if any
+            try:
+                t = t[:t.index(self.tokenizer.eos_id)]
+            except ValueError:
+                pass
+            decoded.append(self.tokenizer.decode(t))
+        return decoded
+
+
+def sample_top_p(probs, p):
+    probs_sort, probs_idx = torch.sort(probs, dim=-1, descending=True)
+    probs_sum = torch.cumsum(probs_sort, dim=-1)
+    mask = probs_sum - probs_sort > p
+    probs_sort[mask] = 0.0
+    probs_sort.div_(probs_sort.sum(dim=-1, keepdim=True))
+    next_token = torch.multinomial(probs_sort, num_samples=1)
+    next_token = torch.gather(probs_idx, -1, next_token)
+    return next_token
+
+
+def load(ckpt_dir: str, tokenizer_path: str, local_rank: int, world_size: int) -> LLaMA:
+    start_time = time.time()
+    checkpoints = sorted(Path(ckpt_dir).glob("*.pth"))
+    assert (world_size == len(checkpoints)
+           ), f"Loading a checkpoint for MP={len(checkpoints)} but world size is {world_size}"
+    ckpt_path = checkpoints[local_rank]
+    print("Loading")
+    checkpoint = torch.load(ckpt_path, map_location="cpu")
+    with open(Path(ckpt_dir) / "params.json", "r") as f:
+        params = json.loads(f.read())
+
+    model_args: LLaMAConfig = LLaMAConfig(max_seq_len=1024, max_batch_size=32, **params)
+    tokenizer = LLaMATokenizer(model_path=tokenizer_path)
+    model_args.vocab_size = tokenizer.n_words
+    torch.set_default_tensor_type(torch.cuda.HalfTensor)
+    model = LLaMA(model_args)
+    torch.set_default_tensor_type(torch.FloatTensor)
+    model.load_state_dict(checkpoint, strict=False)
+
+    generator = LLaMAGenerator(model, tokenizer)
+    print(f"Loaded in {time.time() - start_time:.2f} seconds")
+    return generator

applications/ChatGPT/chatgpt/models/llama/llama_actor.py

@@ -0,0 +1,25 @@
+from typing import Optional
+
+import torch
+
+from ..base import Actor
+from .llama import LLaMA, LLaMAConfig
+
+
+class LLaMAActor(Actor):
+    """
+    BLOOM Actor model.
+
+    Args:
+        pretrained (str): Pretrained model name or path.
+        config (BloomConfig): Model config.
+        checkpoint (bool): Enable gradient checkpointing.
+        lora_rank (int): LoRA rank.
+        lora_train_bias (str): LoRA bias training mode.
+    """
+
+    def __init__(self, config: Optional[LLaMAConfig] = None, lora_rank: int = 0, lora_train_bias: str = 'none') -> None:
+
+        model = LLaMA(config)
+
+        super().__init__(model, lora_rank, lora_train_bias)

applications/ChatGPT/chatgpt/models/llama/llama_critic.py

@@ -0,0 +1,23 @@
+from typing import Optional
+
+import torch.nn as nn
+
+from ..base import Critic
+from .llama import LLaMA, LLaMAConfig
+
+
+class LLaMACritic(Critic):
+    """
+    GPT Critic model.
+
+    Args:
+        pretrained (str): Pretrained model name or path.
+        config (GPT2Config): Model config.
+        checkpoint (bool): Enable gradient checkpointing.
+    """
+
+    def __init__(self, config: Optional[LLaMAConfig] = None, checkpoint: bool = False) -> None:
+
+        model = LLaMA(config)
+        value_head = nn.Linear(model.config.dim, 1)
+        super().__init__(model, value_head)

applications/ChatGPT/chatgpt/models/llama/llama_rm.py

@@ -0,0 +1,23 @@
+from typing import Optional
+
+import torch.nn as nn
+
+from ..base import RewardModel
+from .llama import LLaMA, LLaMAConfig
+
+
+class LLaMARM(RewardModel):
+    """
+    GPT Critic model.
+
+    Args:
+        pretrained (str): Pretrained model name or path.
+        config (GPT2Config): Model config.
+        checkpoint (bool): Enable gradient checkpointing.
+    """
+
+    def __init__(self, config: Optional[LLaMAConfig] = None, checkpoint: bool = False) -> None:
+
+        model = LLaMA(config)
+        value_head = nn.Linear(model.config.dim, 1)
+        super().__init__(model, value_head)