sgl-project--sglang/python/sglang/srt/models/minicpm3.py

# Copyright 2023-2024 SGLang Team
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
# ==============================================================================
"""Inference-only MiniCPM3 model compatible with HuggingFace weights."""

import math
from typing import Any, Dict, Iterable, Optional, Tuple

import torch
from torch import nn
from transformers import PretrainedConfig

from sglang.srt.layers.activation import SiluAndMul
from sglang.srt.layers.layernorm import RMSNorm
from sglang.srt.layers.linear import (
    ColumnParallelLinear,
    MergedColumnParallelLinear,
    ReplicatedLinear,
    RowParallelLinear,
)
from sglang.srt.layers.logits_processor import LogitsProcessor
from sglang.srt.layers.quantization.base_config import QuantizationConfig
from sglang.srt.layers.radix_attention import RadixAttention
from sglang.srt.layers.rotary_embedding import get_rope
from sglang.srt.layers.vocab_parallel_embedding import (
    ParallelLMHead,
    VocabParallelEmbedding,
)
from sglang.srt.model_executor.forward_batch_info import ForwardBatch
from sglang.srt.model_loader.weight_utils import default_weight_loader
from sglang.srt.runtime_context import get_parallel
from sglang.srt.utils import add_prefix, is_cuda
from sglang.srt.utils.hf_transformers_utils import get_rope_config

if is_cuda():
    from sgl_kernel import bmm_fp8 as _raw_bmm_fp8

    from sglang.srt.utils.custom_op import register_custom_op

    # TODO(yuwei): remove this wrapper after sgl-kernel registers its own fake/meta impl
    # Wrap bmm_fp8 as a custom op so torch.compile does not trace into
    # torch.cuda.current_blas_handle() (which returns a non-Tensor).
    @register_custom_op(mutates_args=["out"])
    def _bmm_fp8_op(
        A: torch.Tensor,
        B: torch.Tensor,
        out: torch.Tensor,
        A_scale: torch.Tensor,
        B_scale: torch.Tensor,
    ) -> None:
        _raw_bmm_fp8(A, B, A_scale, B_scale, out.dtype, out)

    def bmm_fp8(A, B, A_scale, B_scale, dtype, out=None):
        if out is None:
            out = torch.empty(
                (A.shape[0], A.shape[1], B.shape[2]),
                device=A.device,
                dtype=dtype,
            )
        _bmm_fp8_op(A, B, out, A_scale, B_scale)
        return out


class MiniCPM3MLP(nn.Module):
    def __init__(
        self,
        hidden_size: int,
        intermediate_size: int,
        hidden_act: str,
        quant_config: Optional[QuantizationConfig] = None,
        prefix: str = "",
    ) -> None:
        super().__init__()
        self.gate_up_proj = MergedColumnParallelLinear(
            hidden_size,
            [intermediate_size] * 2,
            bias=False,
            quant_config=quant_config,
            prefix=add_prefix("gate_up_proj", prefix),
        )
        self.down_proj = RowParallelLinear(
            intermediate_size,
            hidden_size,
            bias=False,
            quant_config=quant_config,
            prefix=add_prefix("down_proj", prefix),
        )
        if hidden_act != "silu":
            raise ValueError(
                f"Unsupported activation: {hidden_act}. "
                "Only silu is supported for now."
            )
        self.act_fn = SiluAndMul()

    def forward(self, x):
        gate_up, _ = self.gate_up_proj(x)
        x = self.act_fn(gate_up)
        x, _ = self.down_proj(x)
        return x


def input_to_float8(x, dtype=torch.float8_e4m3fn):
    finfo = torch.finfo(dtype)
    min_val, max_val = x.aminmax()
    amax = torch.maximum(min_val.abs(), max_val.abs()).clamp(min=1e-12)
    scale = finfo.max / amax
    x_scl_sat = (x * scale).clamp(min=finfo.min, max=finfo.max)
    return x_scl_sat.to(dtype).contiguous(), scale.float().reciprocal()


class MiniCPM3AttentionMLA(nn.Module):

    def __init__(
        self,
        config: PretrainedConfig,
        hidden_size: int,
        num_heads: int,
        qk_nope_head_dim: int,
        qk_rope_head_dim: int,
        v_head_dim: int,
        q_lora_rank: int,
        kv_lora_rank: int,
        rope_theta: float = 10000,
        rope_scaling: Optional[Dict[str, Any]] = None,
        max_position_embeddings: int = 8192,
        quant_config: Optional[QuantizationConfig] = None,
        layer_id=None,
        prefix: str = "",
    ) -> None:
        super().__init__()
        self.layer_id = layer_id
        self.hidden_size = hidden_size
        self.qk_nope_head_dim = qk_nope_head_dim
        self.qk_rope_head_dim = qk_rope_head_dim
        self.qk_head_dim = qk_nope_head_dim + qk_rope_head_dim
        self.v_head_dim = v_head_dim
        self.q_lora_rank = q_lora_rank
        self.kv_lora_rank = kv_lora_rank
        self.num_heads = num_heads
        tp_size = get_parallel().tp_size
        assert num_heads % tp_size == 0
        self.num_local_heads = num_heads // tp_size
        self.scaling = self.qk_head_dim**-0.5
        self.rope_theta = rope_theta
        self.max_position_embeddings = max_position_embeddings

        if self.q_lora_rank is not None:
            self.q_a_proj = ReplicatedLinear(
                self.hidden_size,
                self.q_lora_rank,
                bias=False,
                quant_config=quant_config,
                prefix=add_prefix("q_a_proj", prefix),
            )
            self.q_a_layernorm = RMSNorm(self.q_lora_rank, eps=config.rms_norm_eps)
            self.q_b_proj = ColumnParallelLinear(
                q_lora_rank,
                self.num_heads * self.qk_head_dim,
                bias=False,
                quant_config=quant_config,
                prefix=add_prefix("q_b_proj", prefix),
            )
        else:
            self.q_proj = ColumnParallelLinear(
                self.hidden_size,
                self.num_heads * self.qk_head_dim,
                bias=False,
                quant_config=quant_config,
                prefix=add_prefix("q_proj", prefix),
            )

        self.kv_a_proj_with_mqa = ReplicatedLinear(
            self.hidden_size,
            self.kv_lora_rank + self.qk_rope_head_dim,
            bias=False,
            quant_config=quant_config,
            prefix=add_prefix("kv_a_proj_with_mqa", prefix),
        )
        self.kv_a_layernorm = RMSNorm(self.kv_lora_rank, eps=config.rms_norm_eps)
        self.kv_b_proj = ColumnParallelLinear(
            self.kv_lora_rank,
            self.num_heads * (self.qk_nope_head_dim + self.v_head_dim),
            bias=False,
            quant_config=quant_config,
            prefix=add_prefix("kv_b_proj", prefix),
        )
        # O projection.
        self.o_proj = RowParallelLinear(
            self.num_heads * self.v_head_dim,
            self.hidden_size,
            bias=False,
            quant_config=quant_config,
            prefix=add_prefix("o_proj", prefix),
        )
        self.rotary_emb = get_rope(
            qk_rope_head_dim,
            rotary_dim=qk_rope_head_dim,
            max_position=max_position_embeddings,
            base=rope_theta,
            rope_scaling=rope_scaling,
        )

        self.attn = RadixAttention(
            self.num_local_heads,
            self.kv_lora_rank + self.qk_rope_head_dim,
            self.scaling,
            num_kv_heads=1,
            layer_id=layer_id,
            v_head_dim=self.kv_lora_rank,
            quant_config=quant_config,
            prefix=add_prefix("attn", prefix),
        )

        self.w_kc = None
        self.w_vc = None
        self.w_scale = None

    def forward(
        self,
        positions: torch.Tensor,
        hidden_states: torch.Tensor,
        forward_batch: ForwardBatch,
    ) -> torch.Tensor:
        q_len = hidden_states.shape[0]
        q_input = hidden_states.new_empty(
            q_len, self.num_local_heads, self.kv_lora_rank + self.qk_rope_head_dim
        )
        if self.q_lora_rank is not None:
            q = self.q_a_proj(hidden_states)[0]
            q = self.q_a_layernorm(q)
            q = self.q_b_proj(q)[0].view(-1, self.num_local_heads, self.qk_head_dim)
        else:
            q = self.q_proj(hidden_states)[0].view(
                -1, self.num_local_heads, self.qk_head_dim
            )
        q_nope, q_pe = q.split([self.qk_nope_head_dim, self.qk_rope_head_dim], dim=-1)

        if self.w_kc.dtype == torch.float8_e4m3fn:
            q_nope_val, q_nope_scale = input_to_float8(
                q_nope.transpose(0, 1), torch.float8_e4m3fn
            )
            q_nope_out = bmm_fp8(
                q_nope_val, self.w_kc, q_nope_scale, self.w_scale, torch.bfloat16
            )
        else:
            q_nope_out = torch.bmm(q_nope.transpose(0, 1), self.w_kc)
        q_input[..., : self.kv_lora_rank] = q_nope_out.transpose(0, 1)

        latent_cache = self.kv_a_proj_with_mqa(hidden_states)[0]
        v_input = latent_cache[..., : self.kv_lora_rank]
        v_input = self.kv_a_layernorm(v_input.contiguous()).unsqueeze(1)
        k_input = latent_cache.unsqueeze(1)
        k_input[..., : self.kv_lora_rank] = v_input
        k_pe = k_input[..., self.kv_lora_rank :]

        original_shapes = [q_pe.shape, k_pe.shape]
        q_pe, k_pe = self.rotary_emb(
            positions,
            q_pe.reshape(-1, q_pe.shape[1] * q_pe.shape[2]),
            k_pe.reshape(-1, k_pe.shape[1] * k_pe.shape[2]),
        )
        q_pe, k_pe = q_pe.view(original_shapes[0]), k_pe.view(original_shapes[1])
        q_input[..., self.kv_lora_rank :] = q_pe
        k_input[..., self.kv_lora_rank :] = k_pe

        attn_output = self.attn(q_input, k_input, v_input, forward_batch)
        attn_output = attn_output.view(-1, self.num_local_heads, self.kv_lora_rank)

        if self.w_vc.dtype == torch.float8_e4m3fn:
            attn_output_val, attn_output_scale = input_to_float8(
                attn_output.transpose(0, 1), torch.float8_e4m3fn
            )
            attn_bmm_output = bmm_fp8(
                attn_output_val,
                self.w_vc,
                attn_output_scale,
                self.w_scale,
                torch.bfloat16,
            )
        else:
            attn_bmm_output = torch.bmm(attn_output.transpose(0, 1), self.w_vc)
        attn_output = attn_bmm_output.transpose(0, 1).flatten(1, 2)
        output, _ = self.o_proj(attn_output)

        return output


class MiniCPM3DecoderLayer(nn.Module):
    def __init__(
        self,
        config: PretrainedConfig,
        layer_id: int,
        quant_config: Optional[QuantizationConfig] = None,
        prefix: str = "",
    ) -> None:
        super().__init__()
        self.config = config
        self.hidden_size = config.hidden_size
        rope_theta, rope_scaling = get_rope_config(config)
        max_position_embeddings = getattr(config, "max_position_embeddings", 8192)
        self.self_attn = MiniCPM3AttentionMLA(
            config=config,
            hidden_size=self.hidden_size,
            num_heads=config.num_attention_heads,
            qk_nope_head_dim=config.qk_nope_head_dim,
            qk_rope_head_dim=config.qk_rope_head_dim,
            v_head_dim=self.hidden_size // config.num_attention_heads,
            q_lora_rank=(
                config.q_lora_rank if hasattr(config, "q_lora_rank") else None
            ),
            kv_lora_rank=config.kv_lora_rank,
            rope_theta=rope_theta,
            rope_scaling=rope_scaling,
            max_position_embeddings=max_position_embeddings,
            quant_config=quant_config,
            layer_id=layer_id,
            prefix=add_prefix("self_attn", prefix),
        )

        self.mlp = MiniCPM3MLP(
            hidden_size=self.hidden_size,
            intermediate_size=config.intermediate_size,
            hidden_act=config.hidden_act,
            quant_config=quant_config,
            prefix=add_prefix("mlp", prefix),
        )
        self.input_layernorm = RMSNorm(config.hidden_size, eps=config.rms_norm_eps)
        self.post_attention_layernorm = RMSNorm(
            config.hidden_size, eps=config.rms_norm_eps
        )

    def forward(
        self,
        positions: torch.Tensor,
        hidden_states: torch.Tensor,
        forward_batch: ForwardBatch,
        residual: Optional[torch.Tensor],
    ) -> Tuple[torch.Tensor, torch.Tensor]:
        # Self Attention
        residual = hidden_states
        hidden_states = self.input_layernorm(hidden_states)
        hidden_states = self.self_attn(
            positions=positions,
            hidden_states=hidden_states,
            forward_batch=forward_batch,
        )
        hidden_states = residual + hidden_states * (
            self.config.scale_depth / math.sqrt(self.config.num_hidden_layers)
        )

        # Fully Connected
        residual = hidden_states
        hidden_states = self.post_attention_layernorm(hidden_states)
        hidden_states = self.mlp(hidden_states)
        hidden_states = residual + hidden_states * (
            self.config.scale_depth / math.sqrt(self.config.num_hidden_layers)
        )

        return hidden_states, None


class MiniCPM3Model(nn.Module):
    def __init__(
        self,
        config: PretrainedConfig,
        quant_config: Optional[QuantizationConfig] = None,
        prefix: str = "",
    ) -> None:
        super().__init__()
        self.config = config
        self.padding_idx = config.pad_token_id
        self.vocab_size = config.vocab_size
        self.embed_tokens = VocabParallelEmbedding(
            self.vocab_size,
            config.hidden_size,
            org_num_embeddings=config.vocab_size,
            prefix=add_prefix("embed_tokens", prefix),
        )
        self.layers = nn.ModuleList(
            [
                MiniCPM3DecoderLayer(
                    config,
                    i,
                    quant_config=quant_config,
                    prefix=add_prefix(f"layers.{i}", prefix),
                )
                for i in range(config.num_hidden_layers)
            ]
        )
        self.norm = RMSNorm(config.hidden_size, eps=config.rms_norm_eps)

    def forward(
        self,
        input_ids: torch.Tensor,
        positions: torch.Tensor,
        forward_batch: ForwardBatch,
        input_embeds: torch.Tensor = None,
    ) -> torch.Tensor:
        if input_embeds is None:
            hidden_states = self.embed_tokens(input_ids) * self.config.scale_emb
        else:
            hidden_states = input_embeds
        residual = None

        for i in range(len(self.layers)):
            layer = self.layers[i]
            hidden_states, residual = layer(
                positions,
                hidden_states,
                forward_batch,
                residual,
            )
        hidden_states = self.norm(hidden_states)
        return hidden_states


class MiniCPM3ForCausalLM(nn.Module):
    def __init__(
        self,
        config: PretrainedConfig,
        quant_config: Optional[QuantizationConfig] = None,
        prefix: str = "",
    ) -> None:
        super().__init__()
        self.config = config

        self.num_experts = getattr(self.config, "num_experts", 0)
        self.quant_config = quant_config
        self.model = MiniCPM3Model(
            config, quant_config=quant_config, prefix=add_prefix("model", prefix)
        )
        # self.lm_head = ParallelLMHead(config.vocab_size, config.hidden_size)
        if not self.config.tie_word_embeddings:
            self.lm_head = ParallelLMHead(
                config.vocab_size,
                config.hidden_size,
                org_num_embeddings=config.vocab_size,
                prefix=add_prefix("lm_head", prefix),
            )

        self.scale_width = self.config.hidden_size / self.config.dim_model_base

        self.logits_processor = LogitsProcessor(config)

    @torch.no_grad()
    def forward(
        self,
        input_ids: torch.Tensor,
        positions: torch.Tensor,
        forward_batch: ForwardBatch,
        input_embeds: torch.Tensor = None,
    ) -> torch.Tensor:
        if input_embeds is not None:
            input_embeds = input_embeds * self.config.scale_emb
        hidden_states = self.model(input_ids, positions, forward_batch, input_embeds)
        hidden_states = hidden_states / self.scale_width
        if self.config.tie_word_embeddings:
            lm_head = self.model.embed_tokens
        else:
            lm_head = self.lm_head
        return self.logits_processor(input_ids, hidden_states, lm_head, forward_batch)

    def load_weights(self, weights: Iterable[Tuple[str, torch.Tensor]]):
        stacked_params_mapping = [
            # (param_name, shard_name, shard_id)
            ("gate_up_proj", "gate_proj", 0),
            ("gate_up_proj", "up_proj", 1),
        ]
        expert_params_mapping = [
            # (param_name, weight_name, expert_id)
            (
                "ws" if weight_name in ["w1", "w3"] else "w2s",
                f"experts.{expert_id}.{weight_name}.weight",
                expert_id,
            )
            for expert_id in range(self.num_experts)
            for weight_name in ["w1", "w2", "w3"]
        ]
        params_dict = dict(self.named_parameters())
        for name, loaded_weight in weights:
            if "rotary_emb.inv_freq" in name:
                continue
            if "rotary_emb.cos_cached" in name or "rotary_emb.sin_cached" in name:
                # Models trained using ColossalAI may include these tensors in
                # the checkpoint. Skip them.
                continue
            if self.config.tie_word_embeddings and "lm_head.weight" in name:
                continue

            for param_name, weight_name, shard_id in stacked_params_mapping:
                if weight_name not in name:
                    continue
                name = name.replace(weight_name, param_name)
                # Skip loading extra bias for GPTQ models.
                if name.endswith(".bias") and name not in params_dict:
                    continue
                param = params_dict[name]
                weight_loader = param.weight_loader
                weight_loader(param, loaded_weight, shard_id)
                break
            else:
                for param_name, weight_name, expert_id in expert_params_mapping:
                    if weight_name not in name:
                        continue
                    name = name.replace(weight_name, param_name)
                    param = params_dict[name]
                    weight_loader = param.weight_loader
                    weight_loader(
                        param, loaded_weight, weight_name, expert_id=expert_id
                    )
                    break
                else:
                    # Skip loading extra bias for GPTQ models.
                    if name.endswith(".bias") and name not in params_dict:
                        continue
                    param = params_dict[name]
                    weight_loader = getattr(
                        param, "weight_loader", default_weight_loader
                    )
                    weight_loader(param, loaded_weight)

        for layer_id in range(self.config.num_hidden_layers):
            self_attn = self.model.layers[layer_id].self_attn
            w_kc, w_vc = self_attn.kv_b_proj.weight.unflatten(
                0, (-1, self_attn.qk_nope_head_dim + self_attn.v_head_dim)
            ).split([self_attn.qk_nope_head_dim, self_attn.v_head_dim], dim=1)
            self_attn.w_kc = w_kc.transpose(1, 2).contiguous().transpose(1, 2)
            self_attn.w_vc = w_vc.contiguous().transpose(1, 2)
            if hasattr(self_attn.kv_b_proj, "weight_scale"):
                self_attn.w_scale = self_attn.kv_b_proj.weight_scale
            del self_attn.kv_b_proj


EntryClass = MiniCPM3ForCausalLM