Update custom_bitnet.py

custom_bitnet.py

@@ -1,76 +1,470 @@
-from transformers import PreTrainedModel, PretrainedConfig
+# coding=utf-8
+# Copyright 2025 Microsoft, EleutherAI, and the HuggingFace Inc. team. All rights reserved.
+# Licensed under the Apache License, Version 2.0.
+
+from typing import Optional, Tuple, Union
 import torch
-import torch.nn as nn
+from torch import nn
+from transformers import PreTrainedModel, PretrainedConfig
+from transformers.modeling_outputs import BaseModelOutputWithPast, CausalLMOutputWithPast
+from transformers.cache_utils import DynamicCache
+from transformers.activations import ACT2FN
 
+# BitNetConfig
 class BitNetConfig(PretrainedConfig):
     model_type = "bitnet"
+    keys_to_ignore_at_inference = ["past_key_values"]
+
     def __init__(
         self,
-        vocab_size=32000,
-        hidden_size=768,
-        num_hidden_layers=12,
-        num_attention_heads=12,
-        intermediate_size=3072,
-        hidden_act="gelu",
-        max_position_embeddings=512,
+        vocab_size=128256,
+        hidden_size=2560,
+        intermediate_size=6912,
+        num_hidden_layers=30,
+        num_attention_heads=20,
+        num_key_value_heads=5,
+        hidden_act="relu2",
+        max_position_embeddings=2048,
         initializer_range=0.02,
-        layer_norm_eps=1e-12,
-        dropout=0.1,
-        pad_token_id=0,
-        bos_token_id=1,
-        eos_token_id=2,
-        **kwargs
+        rms_norm_eps=1e-5,
+        use_cache=True,
+        pad_token_id=None,
+        bos_token_id=128000,
+        eos_token_id=128001,
+        tie_word_embeddings=False,
+        rope_theta=500000.0,
+        attention_bias=False,
+        attention_dropout=0.0,
+        **kwargs,
     ):
         self.vocab_size = vocab_size
+        self.max_position_embeddings = max_position_embeddings
         self.hidden_size = hidden_size
+        self.intermediate_size = intermediate_size
         self.num_hidden_layers = num_hidden_layers
         self.num_attention_heads = num_attention_heads
-        self.intermediate_size = intermediate_size
+        self.num_key_value_heads = num_key_value_heads or num_attention_heads
         self.hidden_act = hidden_act
-        self.max_position_embeddings = max_position_embeddings
         self.initializer_range = initializer_range
-        self.layer_norm_eps = layer_norm_eps
-        self.dropout = dropout
+        self.rms_norm_eps = rms_norm_eps
+        self.use_cache = use_cache
+        self.rope_theta = rope_theta
+        self.attention_bias = attention_bias
+        self.attention_dropout = attention_dropout
         super().__init__(
             pad_token_id=pad_token_id,
             bos_token_id=bos_token_id,
             eos_token_id=eos_token_id,
-            **kwargs
+            tie_word_embeddings=tie_word_embeddings,
+            **kwargs,
+        )
+
+# BitNetRMSNorm
+class BitNetRMSNorm(nn.Module):
+    def __init__(self, hidden_size, eps=1e-6):
+        super().__init__()
+        self.weight = nn.Parameter(torch.ones(hidden_size))
+        self.variance_epsilon = eps
+
+    def forward(self, hidden_states):
+        input_dtype = hidden_states.dtype
+        hidden_states = hidden_states.to(torch.float32)
+        variance = hidden_states.pow(2).mean(-1, keepdim=True)
+        hidden_states = hidden_states * torch.rsqrt(variance + self.variance_epsilon)
+        return self.weight * hidden_states.to(input_dtype)
+
+# BitNetMLP
+class BitNetMLP(nn.Module):
+    def __init__(self, config: BitNetConfig):
+        super().__init__()
+        self.config = config
+        self.hidden_size = config.hidden_size
+        self.intermediate_size = config.intermediate_size
+        self.gate_proj = nn.Linear(self.hidden_size, self.intermediate_size, bias=False)
+        self.up_proj = nn.Linear(self.hidden_size, self.intermediate_size, bias=False)
+        self.down_proj = nn.Linear(self.intermediate_size, self.hidden_size, bias=False)
+        self.act_fn = ACT2FN[config.hidden_act]
+        self.ffn_sub_norm = BitNetRMSNorm(config.intermediate_size, eps=config.rms_norm_eps)
+
+    def forward(self, x):
+        down_proj = self.down_proj(self.ffn_sub_norm(self.act_fn(self.gate_proj(x)) * self.up_proj(x)))
+        return down_proj
+
+# Utility Functions
+def rotate_half(x):
+    x1 = x[..., : x.shape[-1] // 2]
+    x2 = x[..., x.shape[-1] // 2 :]
+    return torch.cat((-x2, x1), dim=-1)
+
+def apply_rotary_pos_emb(q, k, cos, sin, position_ids=None, unsqueeze_dim=1):
+    cos = cos.unsqueeze(unsqueeze_dim)
+    sin = sin.unsqueeze(unsqueeze_dim)
+    q_embed = (q * cos) + (rotate_half(q) * sin)
+    k_embed = (k * cos) + (rotate_half(k) * sin)
+    return q_embed, k_embed
+
+def repeat_kv(hidden_states: torch.Tensor, n_rep: int) -> torch.Tensor:
+    batch, num_key_value_heads, slen, head_dim = hidden_states.shape
+    if n_rep == 1:
+        return hidden_states
+    hidden_states = hidden_states[:, :, None, :, :].expand(batch, num_key_value_heads, n_rep, slen, head_dim)
+    return hidden_states.reshape(batch, num_key_value_heads * n_rep, slen, head_dim)
+
+def eager_attention_forward(
+    module: nn.Module,
+    query: torch.Tensor,
+    key: torch.Tensor,
+    value: torch.Tensor,
+    attention_mask: Optional[torch.Tensor],
+    scaling: float,
+    dropout: float = 0.0,
+    **kwargs,
+):
+    key_states = repeat_kv(key, module.num_key_value_groups)
+    value_states = repeat_kv(value, module.num_key_value_groups)
+    attn_weights = torch.matmul(query, key_states.transpose(2, 3)) * scaling
+    if attention_mask is not None:
+        causal_mask = attention_mask[:, :, :, : key_states.shape[-2]]
+        attn_weights = attn_weights + causal_mask
+    attn_weights = nn.functional.softmax(attn_weights, dim=-1, dtype=torch.float32).to(query.dtype)
+    attn_weights = nn.functional.dropout(attn_weights, p=dropout, training=module.training)
+    attn_output = torch.matmul(attn_weights, value_states)
+    attn_output = attn_output.transpose(1, 2).contiguous()
+    return attn_output, attn_weights
+
+# BitNetAttention
+class BitNetAttention(nn.Module):
+    def __init__(self, config: BitNetConfig, layer_idx: int):
+        super().__init__()
+        self.config = config
+        self.layer_idx = layer_idx
+        self.head_dim = config.hidden_size // config.num_attention_heads
+        self.num_key_value_groups = config.num_attention_heads // config.num_key_value_heads
+        self.scaling = self.head_dim**-0.5
+        self.attention_dropout = config.attention_dropout
+        self.is_causal = True
+        self.q_proj = nn.Linear(
+            config.hidden_size, config.num_attention_heads * self.head_dim, bias=config.attention_bias
+        )
+        self.k_proj = nn.Linear(
+            config.hidden_size, config.num_key_value_heads * self.head_dim, bias=config.attention_bias
+        )
+        self.v_proj = nn.Linear(
+            config.hidden_size, config.num_key_value_heads * self.head_dim, bias=config.attention_bias
+        )
+        self.o_proj = nn.Linear(
+            config.num_attention_heads * self.head_dim, config.hidden_size, bias=config.attention_bias
+        )
+        self.attn_sub_norm = BitNetRMSNorm(config.hidden_size, eps=config.rms_norm_eps)
+
+    def forward(
+        self,
+        hidden_states: torch.Tensor,
+        position_embeddings: Tuple[torch.Tensor, torch.Tensor],
+        attention_mask: Optional[torch.Tensor],
+        past_key_value: Optional[DynamicCache] = None,
+        cache_position: Optional[torch.LongTensor] = None,
+        **kwargs,
+    ) -> Tuple[torch.Tensor, Optional[torch.Tensor], Optional[Tuple[torch.Tensor]]]:
+        input_shape = hidden_states.shape[:-1]
+        hidden_shape = (*input_shape, -1, self.head_dim)
+        query_states = self.q_proj(hidden_states).view(hidden_shape).transpose(1, 2)
+        key_states = self.k_proj(hidden_states).view(hidden_shape).transpose(1, 2)
+        value_states = self.v_proj(hidden_states).view(hidden_shape).transpose(1, 2)
+        cos, sin = position_embeddings
+        query_states, key_states = apply_rotary_pos_emb(query_states, key_states, cos, sin)
+        if past_key_value is not None:
+            cache_kwargs = {"sin": sin, "cos": cos, "cache_position": cache_position}
+            key_states, value_states = past_key_value.update(key_states, value_states, self.layer_idx, cache_kwargs)
+        attn_output, attn_weights = eager_attention_forward(
+            self,
+            query_states,
+            key_states,
+            value_states,
+            attention_mask,
+            dropout=0.0 if not self.training else self.attention_dropout,
+            scaling=self.scaling,
+        )
+        attn_output = attn_output.reshape(*input_shape, -1).contiguous()
+        attn_output = self.attn_sub_norm(attn_output)
+        attn_output = self.o_proj(attn_output)
+        return attn_output, attn_weights
+
+# BitNetDecoderLayer
+class BitNetDecoderLayer(nn.Module):
+    def __init__(self, config: BitNetConfig, layer_idx: int):
+        super().__init__()
+        self.hidden_size = config.hidden_size
+        self.self_attn = BitNetAttention(config=config, layer_idx=layer_idx)
+        self.mlp = BitNetMLP(config)
+        self.input_layernorm = BitNetRMSNorm(config.hidden_size, eps=config.rms_norm_eps)
+        self.post_attention_layernorm = BitNetRMSNorm(config.hidden_size, eps=config.rms_norm_eps)
+
+    def forward(
+        self,
+        hidden_states: torch.Tensor,
+        attention_mask: Optional[torch.Tensor] = None,
+        position_ids: Optional[torch.LongTensor] = None,
+        past_key_value: Optional[DynamicCache] = None,
+        output_attentions: Optional[bool] = False,
+        use_cache: Optional[bool] = False,
+        cache_position: Optional[torch.LongTensor] = None,
+        position_embeddings: Optional[Tuple[torch.Tensor, torch.Tensor]] = None,
+        **kwargs,
+    ) -> Tuple[torch.FloatTensor, Optional[Tuple[torch.FloatTensor, torch.FloatTensor]]]:
+        residual = hidden_states
+        hidden_states = self.input_layernorm(hidden_states)
+        hidden_states, self_attn_weights = self.self_attn(
+            hidden_states=hidden_states,
+            attention_mask=attention_mask,
+            position_ids=position_ids,
+            past_key_value=past_key_value,
+            output_attentions=output_attentions,
+            use_cache=use_cache,
+            cache_position=cache_position,
+            position_embeddings=position_embeddings,
+            **kwargs,
+        )
+        hidden_states = residual + hidden_states
+        residual = hidden_states
+        hidden_states = self.post_attention_layernorm(hidden_states)
+        hidden_states = self.mlp(hidden_states)
+        hidden_states = residual + hidden_states
+        outputs = (hidden_states,)
+        if output_attentions:
+            outputs += (self_attn_weights,)
+        return outputs
+
+# BitNetRotaryEmbedding
+class BitNetRotaryEmbedding(nn.Module):
+    def __init__(self, config: BitNetConfig, device=None):
+        super().__init__()
+        self.rope_type = "default"
+        self.max_seq_len_cached = config.max_position_embeddings
+        self.config = config
+        dim = config.hidden_size // config.num_attention_heads
+        inv_freq = 1.0 / (config.rope_theta ** (torch.arange(0, dim, 2, dtype=torch.float, device=device) / dim))
+        self.register_buffer("inv_freq", inv_freq, persistent=False)
+        self.original_inv_freq = self.inv_freq
+
+    def forward(self, x, position_ids):
+        inv_freq_expanded = self.inv_freq[None, :, None].float().expand(position_ids.shape[0], -1, 1).to(x.device)
+        position_ids_expanded = position_ids[:, None, :].float()
+        device_type = x.device.type if isinstance(x.device.type, str) and x.device.type != "mps" else "cpu"
+        with torch.autocast(device_type=device_type, enabled=False):
+            freqs = (inv_freq_expanded.float() @ position_ids_expanded.float()).transpose(1, 2)
+            emb = torch.cat((freqs, freqs), dim=-1)
+            cos = emb.cos()
+            sin = emb.sin()
+        return cos.to(dtype=x.dtype), sin.to(dtype=x.dtype)
+
+# BitNetModel
+class BitNetModel(PreTrainedModel):
+    config_class = BitNetConfig
+    supports_gradient_checkpointing = True
+    _no_split_modules = ["BitNetDecoderLayer"]
+
+    def __init__(self, config: BitNetConfig):
+        super().__init__(config)
+        self.padding_idx = config.pad_token_id
+        self.vocab_size = config.vocab_size
+        self.embed_tokens = nn.Embedding(config.vocab_size, config.hidden_size, self.padding_idx)
+        self.layers = nn.ModuleList(
+            [BitNetDecoderLayer(config, layer_idx) for layer_idx in range(config.num_hidden_layers)]
+        )
+        self.norm = BitNetRMSNorm(config.hidden_size, eps=config.rms_norm_eps)
+        self.rotary_emb = BitNetRotaryEmbedding(config=config)
+        self.gradient_checkpointing = False
+        self.post_init()
+
+    def get_input_embeddings(self):
+        return self.embed_tokens
+
+    def set_input_embeddings(self, value):
+        self.embed_tokens = value
+
+    def forward(
+        self,
+        input_ids: Optional[torch.LongTensor] = None,
+        attention_mask: Optional[torch.Tensor] = None,
+        position_ids: Optional[torch.LongTensor] = None,
+        past_key_values: Optional[DynamicCache] = None,
+        inputs_embeds: Optional[torch.FloatTensor] = None,
+        use_cache: Optional[bool] = None,
+        output_attentions: Optional[bool] = None,
+        output_hidden_states: Optional[bool] = None,
+        cache_position: Optional[torch.LongTensor] = None,
+        **kwargs,
+    ) -> BaseModelOutputWithPast:
+        output_attentions = output_attentions if output_attentions is not None else self.config.output_attentions
+        output_hidden_states = (
+            output_hidden_states if output_hidden_states is not None else self.config.output_hidden_states
+        )
+        use_cache = use_cache if use_cache is not None else self.config.use_cache
+        if (input_ids is None) ^ (inputs_embeds is not None):
+            raise ValueError("You must specify exactly one of input_ids or inputs_embeds")
+        if inputs_embeds is None:
+            inputs_embeds = self.embed_tokens(input_ids)
+        if use_cache and past_key_values is None:
+            past_key_values = DynamicCache()
+        if cache_position is None:
+            past_seen_tokens = past_key_values.get_seq_length() if past_key_values is not None else 0
+            cache_position = torch.arange(
+                past_seen_tokens, past_seen_tokens + inputs_embeds.shape[1], device=inputs_embeds.device
+            )
+        if position_ids is None:
+            position_ids = cache_position.unsqueeze(0)
+        causal_mask = self._update_causal_mask(attention_mask, inputs_embeds, cache_position, past_key_values)
+        hidden_states k= inputs_embeds
+        position_embeddings = self.rotary_emb(hidden_states, position_ids)
+        all_hidden_states = () if output_hidden_states else None
+        all_self_attns = () if output_attentions else None
+        for decoder_layer in self.layers:
+            if output_hidden_states:
+                all_hidden_states += (hidden_states,)
+            layer_outputs = decoder_layer(
+                hidden_states,
+                attention_mask=causal_mask,
+                position_ids=position_ids,
+                past_key_value=past_key_values,
+                output_attentions=output_attentions,
+                use_cache=use_cache,
+                cache_position=cache_position,
+                position_embeddings=position_embeddings,
+            )
+            hidden_states = layer_outputs[0]
+            if output_attentions:
+                all_self_attns += (layer_outputs[1],)
+        hidden_states = self.norm(hidden_states)
+        if output_hidden_states:
+            all_hidden_states += (hidden_states,)
+        return BaseModelOutputWithPast(
+            last_hidden_state=hidden_states,
+            past_key_values=past_key_values if use_cache else None,
+            hidden_states=all_hidden_states,
+            attentions=all_self_attns,
+        )
+
+    def _update_causal_mask(
+        self,
+        attention_mask: Optional[torch.Tensor],
+        input_tensor: torch.Tensor,
+        cache_position: torch.Tensor,
+        past_key_values: Optional[DynamicCache],
+    ):
+        dtype, device = input_tensor.dtype, input_tensor.device
+        sequence_length = input_tensor.shape[1]
+        target_length = past_key_values.get_seq_length() + sequence_length + 1 if past_key_values else sequence_length + 1
+        min_dtype = torch.finfo(dtype).min
+        causal_mask = torch.full(
+            (sequence_length, target_length), fill_value=min_dtype, dtype=dtype, device=device
         )
+        if sequence_length != 1:
+            causal_mask = torch.triu(causal_mask, diagonal=1)
+        causal_mask *= torch.arange(target_length, device=device) > cache_position.reshape(-1, 1)
+        causal_mask = causal_mask[None, None, :, :].expand(input_tensor.shape[0], 1, -1, -1)
+        if attention_mask is not None:
+            causal_mask = causal_mask.clone()
+            mask_length = attention_mask.shape[-1]
+            padding_mask = causal_mask[:, :, :, :mask_length] + attention_mask[:, None, None, :].to(causal_mask.device)
+            padding_mask = padding_mask == 0
+            causal_mask[:, :, :, :mask_length] = causal_mask[:, :, :, :mask_length].masked_fill(padding_mask, min_dtype)
+        return causal_mask
 
+# BitNetForCausalLM
 class BitNetForCausalLM(PreTrainedModel):
     config_class = BitNetConfig
+    _tied_weights_keys = ["lm_head.weight"]
+
     def __init__(self, config):
         super().__init__(config)
-        self.embed_tokens = nn.Embedding(config.vocab_size, config.hidden_size)
-        self.layers = nn.ModuleList([
-            nn.TransformerEncoderLayer(
-                d_model=config.hidden_size,
-                nhead=config.num_attention_heads,
-                dim_feedforward=config.intermediate_size,
-                dropout=config.dropout
-            ) for _ in range(config.num_hidden_layers)
-        ])
-        self.norm = nn.LayerNorm(config.hidden_size, eps=config.layer_norm_eps)
+        self.model = BitNetModel(config)
+        self.vocab_size = config.vocab_size
         self.lm_head = nn.Linear(config.hidden_size, config.vocab_size, bias=False)
-        self.apply(self._init_weights)
-    def _init_weights(self, module):
-        if isinstance(module, nn.Linear):
-            torch.nn.init.normal_(module.weight, mean=0.0, std=self.config.initializer_range)
-            if module.bias is not None:
-                torch.nn.init.zeros_(module.bias)
-        elif isinstance(module, nn.Embedding):
-            torch.nn.init.normal_(module.weight, mean=0.0, std=self.config.initializer_range)
-    def forward(self, input_ids, attention_mask=None, labels=None, **kwargs):
-        hidden_states = self.embed_tokens(input_ids)
-        for layer in self.layers:
-            hidden_states = layer(hidden_states)
-        hidden_states = self.norm(hidden_states)
+        self.post_init()
+
+    def get_input_embeddings(self):
+        return self.model.embed_tokens
+
+    def set_input_embeddings(self, value):
+        self.model.embed_tokens = value
+
+    def get_output_embeddings(self):
+        return self.lm_head
+
+    def set_output_embeddings(self, new_embeddings):
+        self.lm_head = new_embeddings
+
+    def set_decoder(self, decoder):
+        self.model = decoder
+
+    def get_decoder(self):
+        return self.model
+
+    def forward(
+        self,
+        input_ids: Optional[torch.LongTensor] = None,
+        attention_mask: Optional[torch.Tensor] = None,
+        position_ids: Optional[torch.LongTensor] = None,
+        past_key_values: Optional[DynamicCache] = None,
+        inputs_embeds: Optional[torch.FloatTensor] = None,
+        labels: Optional[torch.LongTensor] = None,
+        use_cache: Optional[bool] = None,
+        output_attentions: Optional[bool] = None,
+        output_hidden_states: Optional[bool] = None,
+        cache_position: Optional[torch.LongTensor] = None,
+        **kwargs,
+    ) -> CausalLMOutputWithPast:
+        output_attentions = output_attentions if output_attentions is not None else self.config.output_attentions
+        output_hidden_states = (
+            output_hidden_states if output_hidden_states is not None else self.config.output_hidden_states
+        )
+        outputs = self.model(
+            input_ids=input_ids,
+            attention_mask=attention_mask,
+            position_ids=position_ids,
+            past_key_values=past_key_values,
+            inputs_embeds=inputs_embeds,
+            use_cache=use_cache,
+            output_attentions=output_attentions,
+            output_hidden_states=output_hidden_states,
+            cache_position=cache_position,
+            **kwargs,
+        )
+        hidden_states = outputs.last_hidden_state
         logits = self.lm_head(hidden_states)
         loss = None
         if labels is not None:
             loss_fct = nn.CrossEntropyLoss()
             loss = loss_fct(logits.view(-1, self.config.vocab_size), labels.view(-1))
-        return {"logits": logits, "loss": loss} if loss is not None else {"logits": logits}
-    def prepare_inputs_for_generation(self, input_ids, **kwargs):
-        return {"input_ids": input_ids, **kwargs}
+        return CausalLMOutputWithPast(
+            loss=loss,
+            logits=logits,
+            past_key_values=outputs.past_key_values,
+            hidden_states=outputs.hidden_states,
+            attentions=outputs.attentions,
+        )
+
+    def prepare_inputs_for_generation(self, input_ids, past_key_values=None, attention_mask=None, **kwargs):
+        if past_key_values is None:
+            past_key_values = DynamicCache()
+        cache_position = kwargs.get("cache_position", None)
+        if cache_position is None:
+            past_length = past_key_values.get_seq_length() if past_key_values is not None else 0
+            cache_position = torch.arange(past_length, past_length + input_ids.shape[-1], device=input_ids.device)
+        position_ids = cache_position.unsqueeze(0)
+        if attention_mask is not None and attention_mask.shape[1] != input_ids.shape[1]:
+            attention_mask = self._update_causal_mask(
+                attention_mask,
+                input_ids,
+                cache_position,
+                past_key_values
+            )
+        return {
+            "input_ids": input_ids,
+            "position_ids": position_ids,
+            "attention_mask": attention_mask,
+            "past_key_values": past_key_values,
+            "cache_position": cache_position,
+            "use_cache": kwargs.get("use_cache", True),
+        }