From 4e1236cbf64b6bcee543732c7d89247d457dc947 Mon Sep 17 00:00:00 2001 From: Goekdeniz-Guelmez Date: Fri, 11 Oct 2024 20:53:29 +0200 Subject: [PATCH] fixing loading the model --- llms/mamba2-130m-hf | 1 + llms/mlx_lm/models/mamba2 copy.py | 256 +++++++++++++++++++++++++++ llms/mlx_lm/models/mamba2-other.py | 275 +++++++++++++++++++++++++++++ llms/mlx_lm/models/mamba2.py | 209 ++++++++++++---------- 4 files changed, 644 insertions(+), 97 deletions(-) create mode 160000 llms/mamba2-130m-hf create mode 100644 llms/mlx_lm/models/mamba2 copy.py create mode 100644 llms/mlx_lm/models/mamba2-other.py diff --git a/llms/mamba2-130m-hf b/llms/mamba2-130m-hf new file mode 160000 index 00000000..05e8773f --- /dev/null +++ b/llms/mamba2-130m-hf @@ -0,0 +1 @@ +Subproject commit 05e8773fc4ac1cd067e8a18a5c45372ce5178405 diff --git a/llms/mlx_lm/models/mamba2 copy.py b/llms/mlx_lm/models/mamba2 copy.py new file mode 100644 index 00000000..5ecbb3d2 --- /dev/null +++ b/llms/mlx_lm/models/mamba2 copy.py @@ -0,0 +1,256 @@ +# Copyright © 2024 Apple Inc. + +import math +from dataclasses import dataclass, field +from typing import Tuple, Union + +import mlx.core as mx +import mlx.nn as nn + +from .base import BaseModelArgs + + +@dataclass +class ModelArgs(BaseModelArgs): + model_type: str = "mamba2" + num_heads: int = 128 + head_dim: int = 64 + vocab_size: int = 32768 + hidden_size: int = 4096 + state_size: int = 128 + num_hidden_layers: int = 64 + layer_norm_epsilon: float = 1e-5 + pad_token_id: int = 1 + bos_token_id: int = 0 + eos_token_id: int = 2 + expand: int = 2 + conv_kernel: int = 4 + n_groups: int = 8 + use_bias: bool = False + use_conv_bias: bool = True + hidden_act: str = "silu" + initializer_range: float = 0.1 + residual_in_fp32: bool = True + time_step_rank: Union[int, str] = "auto" + time_step_min: float = 0.001 + time_step_max: float = 0.1 + time_step_floor: float = 1e-4 + time_step_limit: Tuple[float, float] = field(default_factory=lambda: (0.0, float("inf"))) + rescale_prenorm_residual: bool = False + use_cache: bool = True + rms_norm: bool = True + chunk_size: int = 256 + tie_word_embeddings: bool = False + + def __post_init__(self): + if not hasattr(self, "intermediate_size"): + self.intermediate_size = int(self.expand * self.hidden_size) + if not hasattr(self, "head_dim"): + self.head_dim = self.hidden_size // self.num_heads + if self.time_step_rank == "auto": + self.time_step_rank = math.ceil(self.hidden_size / 16) + + +class Mamba2Cache: + def __init__(self): + self.cache = [None, None] + + def __setitem__(self, idx, value): + self.cache[idx] = value + + def __getitem__(self, idx): + return self.cache[idx] + + @property + def state(self): + return self.cache + + +class MambaRMSNormGated(nn.Module): + def __init__(self, hidden_size, eps=1e-6): + super().__init__() + self.weight = mx.ones((hidden_size,)) + self.variance_epsilon = eps + + def __call__(self, hidden_states, gate=None): + if gate is not None: + hidden_states = hidden_states * nn.silu(gate) + variance = mx.mean(hidden_states ** 2, axis=-1, keepdims=True) + hidden_states = hidden_states * mx.rsqrt(variance + self.variance_epsilon) + return self.weight * hidden_states + +class DepthWiseConv1d(nn.Module): + def __init__(self, channels, kernel_size, bias=True, groups=1, padding=0): + super().__init__() + self.channels = channels + self.kernel_size = kernel_size + self.padding = padding + self.groups = groups + self.weight = mx.random.normal((self.channels, kernel_size, 1)) + self.bias = mx.zeros((channels,)) if bias else None + + def __call__(self, x, cache=None): + B, L, C = x.shape + _, K, _ = self.weight.shape + + if cache is not None: + x = mx.concatenate([cache, x], axis=1) + else: + x = mx.pad(x, [(0, 0), (K - 1, 0), (0, 0)]) + + y = mx.conv_general(x, self.weight, groups=self.groups) + + if self.bias is not None: + y = y + self.bias + + return y, x[:, -K + 1 :, :] + + +class Mamba2Mixer(nn.Module): + def __init__(self, args: ModelArgs): + super().__init__() + self.args = args + self.intermediate_size = args.intermediate_size + self.time_step_rank = args.time_step_rank + self.conv_kernel_size = args.conv_kernel + self.hidden_size = args.hidden_size + self.state_size = args.state_size + self.num_heads = args.num_heads + self.head_dim = args.head_dim + self.n_groups = args.n_groups + + self.conv_dim = self.intermediate_size + 2 * self.n_groups * self.state_size + self.conv1d = DepthWiseConv1d( + channels=self.conv_dim, + kernel_size=self.conv_kernel_size, + bias=self.args.use_conv_bias, + groups=self.conv_dim, + padding=self.conv_kernel_size - 1, + ) + + projection_size = self.intermediate_size + self.conv_dim + self.num_heads + self.in_proj = nn.Linear( + self.hidden_size, + projection_size, + bias=args.use_bias + ) + + self.act = nn.SiLU() + self.dt_bias = mx.ones((self.num_heads,)) + self.A_log = mx.log(mx.arange(1, self.num_heads + 1)) + self.D = mx.ones((self.num_heads,)) + + self.norm = MambaRMSNormGated(self.intermediate_size, eps=args.layer_norm_epsilon) + + self.out_proj = nn.Linear(self.intermediate_size, self.hidden_size, bias=args.use_bias) + + def ssm_step(self, x, state=None): + A = -mx.exp(self.A_log) + D = self.D + deltaBC = self.x_proj(x) + delta, B, C = mx.split( + deltaBC, + indices_or_sections=[ + self.time_step_rank, + self.time_step_rank + self.ssm_state_size, + ], + axis=-1, + ) + delta = nn.softplus(self.dt_proj(delta)) + new_state = mx.expand_dims(delta * x, -1) * mx.expand_dims(B, 1) + if state is not None: + new_state += state * mx.exp(mx.expand_dims(delta, -1) * A) + y = (new_state @ mx.expand_dims(C, -1)).squeeze(2) + y = y + D * x + return y, new_state + + def __call__(self, x, cache): + B, T, D = x.shape + if cache is None: + cache = [None, None] + + outputs = [] + for t in range(T): + xt = x[:, t, :] + xz = self.in_proj(xt) + x_t, z_t = xz.split(indices_or_sections=2, axis=1) + conv_out, cache[0] = self.conv1d(mx.expand_dims(x_t, 1), cache[0]) + x_t = conv_out.squeeze(1) + x_t = nn.silu(x_t) + y_t, cache[1] = self.ssm_step(x_t, cache[1]) + z_t = nn.silu(z_t) + output_t = y_t * z_t + output_t = self.out_proj(output_t) + outputs.append(output_t) + output = mx.stack(outputs, axis=1) + return output + + +class Mamba2Block(nn.Module): + def __init__(self, args: ModelArgs): + super().__init__() + self.mixer = Mamba2Mixer(args) + self.norm = nn.RMSNorm(args.hidden_size) + + def __call__(self, x: mx.array, cache): + return self.mixer(self.norm(x), cache) + x + + +class Mamba2(nn.Module): + def __init__(self, args: ModelArgs): + super().__init__() + self.args = args + self.embeddings = nn.Embedding(args.vocab_size, args.hidden_size) + self.layers = [Mamba2Block(args) for idx in range(args.num_hidden_layers)] + self.norm_f = nn.RMSNorm(args.hidden_size, eps=args.layer_norm_epsilon) + + def __call__( + self, + inputs: mx.array, + cache=None + ): + hidden_states = self.embeddings(inputs) + + if cache is None: + cache = Mamba2Cache(len(self.layers)) + + for i, layer in enumerate(self.layers): + hidden_states = layer(hidden_states, cache[i]) + + hidden_states = self.norm_f(hidden_states) + return hidden_states + + +class Model(nn.Module): + def __init__(self, args: ModelArgs): + super().__init__() + self.args = args + self.model_type = args.model_type + self.backbone = Mamba2(args) + if not args.tie_word_embeddings: + self.lm_head = nn.Linear(args.hidden_size, args.vocab_size, bias=False) + + def __call__(self, inputs: mx.array, cache=None): + B, T = inputs.shape + + x = self.backbone(inputs, cache) + + if self.args.tie_word_embeddings: + logits = self.backbone.embeddings.as_linear(x) + else: + logits = self.lm_head(x) + + return logits + + def sanitize(self, weights): + for k, v in weights.items(): + if "conv1d.weight" in k and v.ndim == 3: + weights[k] = v.moveaxis(2, 1) + return weights + + def make_cache(self, batch_size: int = 1): + return [Mamba2Cache() for _ in range(len(self.layers))] + + @property + def layers(self): + return self.backbone.layers diff --git a/llms/mlx_lm/models/mamba2-other.py b/llms/mlx_lm/models/mamba2-other.py new file mode 100644 index 00000000..922842b0 --- /dev/null +++ b/llms/mlx_lm/models/mamba2-other.py @@ -0,0 +1,275 @@ +# Copyright © 2024 Apple Inc. + +import math +from dataclasses import dataclass, field +from typing import Optional, Tuple, Union + +import mlx.core as mx +import mlx.nn as nn + +from .base import BaseModelArgs + + +@dataclass +class ModelArgs(BaseModelArgs): + model_type: str = "mamba2" + num_heads: int = 128 + head_dim: int = 64 + vocab_size: int = 32768 + hidden_size: int = 4096 + state_size: int = 128 + num_hidden_layers: int = 64 + layer_norm_epsilon: float = 1e-5 + expand: int = 2 + conv_kernel: int = 4 + n_groups: int = 8 + use_bias: bool = False + use_conv_bias: bool = True + initializer_range: float = 0.1 + residual_in_fp32: bool = True + time_step_rank: Union[int, str] = "auto" + time_step_min: float = 0.001 + time_step_max: float = 0.1 + time_step_floor: float = 1e-4 + time_step_limit: Tuple[float, float] = field(default_factory=lambda: (0.0, float("inf"))) + rescale_prenorm_residual: bool = False + use_cache: bool = True + rms_norm: bool = True + chunk_size: int = 256 + tie_word_embeddings: bool = False + + def __post_init__(self): + if not hasattr(self, "intermediate_size"): + self.intermediate_size = int(self.expand * self.hidden_size) + if not hasattr(self, "head_dim"): + self.head_dim = self.hidden_size // self.num_heads + if self.time_step_rank == "auto": + self.time_step_rank = math.ceil(self.hidden_size / 16) + + +class Mamba2Cache: + def __init__(self, num_layers): + self.cache = [[None, None] for _ in range(num_layers)] + + def __getitem__(self, idx): + return self.cache[idx] + + def __setitem__(self, idx, value): + self.cache[idx] = value + + +class MambaRMSNormGated(nn.Module): + def __init__(self, hidden_size, eps=1e-6): + super().__init__() + self.weight = mx.ones((hidden_size,)) + self.variance_epsilon = eps + + def __call__(self, hidden_states, gate=None): + if gate is not None: + hidden_states = hidden_states * nn.silu(gate) + variance = mx.mean(hidden_states ** 2, axis=-1, keepdims=True) + hidden_states = hidden_states * mx.rsqrt(variance + self.variance_epsilon) + return self.weight * hidden_states + + +class Mamba2Mixer(nn.Module): + def __init__(self, args: ModelArgs): + super().__init__() + self.args = args + self.intermediate_size = args.intermediate_size + self.time_step_rank = args.time_step_rank + self.conv_kernel_size = args.conv_kernel + self.hidden_size = args.hidden_size + self.state_size = args.state_size + self.num_heads = args.num_heads + self.head_dim = args.head_dim + self.n_groups = args.n_groups + + self.conv_dim = self.intermediate_size + 2 * self.n_groups * self.state_size + self.conv1d = nn.Conv1d( + in_channels=self.conv_dim, + out_channels=self.conv_dim, + bias=args.use_conv_bias, + kernel_size=args.conv_kernel, + groups=self.conv_dim, + padding=args.conv_kernel - 1 + ) + + projection_size = self.intermediate_size + self.conv_dim + self.num_heads + self.in_proj = nn.Linear( + self.hidden_size, + projection_size, + bias=args.use_bias + ) + + self.act = nn.SiLU() + self.dt_bias = mx.ones((self.num_heads,)) + self.A_log = mx.log(mx.arange(1, self.num_heads + 1)) + self.D = mx.ones((self.num_heads,)) + + self.norm = MambaRMSNormGated(self.intermediate_size, eps=args.layer_norm_epsilon) + + self.out_proj = nn.Linear(self.intermediate_size, self.hidden_size, bias=args.use_bias) + + def ssm_step(self, x, dt, state): + B, L, C = x.shape + print(f"x shape: {x.shape}") + projected_states = self.in_proj(x) + print(f"deltaBC shape: {projected_states.shape}") + + d_mlp = (projected_states.shape[-1] - 2 * self.intermediate_size - 2 * self.n_groups * self.state_size - self.num_heads) // 2 + + gate = projected_states[:, :, 2*d_mlp:2*d_mlp+self.intermediate_size] + conv_state = projected_states[:, :, 2*d_mlp+self.intermediate_size:2*d_mlp+self.intermediate_size+self.conv_dim] + time_step = projected_states[:, :, -self.num_heads:] + + print(f"conv_state shape before reshape: {conv_state.shape}") + print(f"self.conv_dim: {self.conv_dim}") + + # Reshape and handle the case where L=1 + conv_state = conv_state.reshape(B, self.conv_dim, L) + if L == 1: + # If sequence length is 1, we need to pad to apply convolution + conv_state = mx.pad(conv_state, ((0, 0), (0, 0), (0, self.conv_kernel_size - 1))) + + conv_out = self.conv1d(conv_state) + + # If we padded, we need to remove the padding + if L == 1: + conv_out = conv_out[:, :, :L] + + # Reshape back to (B, L, C) + conv_out = conv_out.transpose(0, 2, 1) + + x_and_conv_out, B, C = mx.split( + conv_out, + [self.intermediate_size, self.n_groups * self.state_size], + axis=-1 + ) + + dt = nn.softplus(time_step + self.dt_bias) + dt = mx.clip(dt, self.args.time_step_min, self.args.time_step_max) + + B = B.reshape(-1, self.num_heads, self.head_dim, self.state_size) + C = C.reshape(-1, self.num_heads, self.head_dim, self.state_size) + + dA = mx.exp(dt[:, :, None, None] * A[None, :, None, None]) + dB = dt[:, :, None, None] * B + + new_state = state * dA + x_and_conv_out[:, :, None, None] * dB + y = mx.sum(new_state * C, axis=-1) + y = y + C[None, :, None] * x_and_conv_out + + y = self.norm(y.reshape(-1, self.intermediate_size), gate) + output = self.out_proj(y) + + return output, new_state + + def __call__( + self, + x: mx.array, + cache = None + ): + B, L, _ = x.shape + + if cache[0] is not None: # Using cached state + conv_state, ssm_state = cache + x = x[:, -1:] + output, new_ssm_state = self.ssm_step(x, None, ssm_state) + cache[1] = new_ssm_state # Update SSM state in cache + else: + conv_state, ssm_state = None, None + outputs = [] + for t in range(L): + x = x[:, t:t+1] + output, ssm_state = self.ssm_step(x, None, ssm_state) + outputs.append(output) + output = mx.concatenate(outputs, axis=1) + cache[1] = ssm_state # Store final SSM state in cache + + # Update conv state in cache + new_conv_state = x[:, -self.conv_kernel_size:] + cache[0] = new_conv_state + + return output + + +class Mamba2Block(nn.Module): + def __init__(self, args: ModelArgs): + super().__init__() + self.args = args + self.residual_in_fp32 = args.residual_in_fp32 + self.norm = nn.RMSNorm(args.hidden_size, eps=args.layer_norm_epsilon) + self.mixer = Mamba2Mixer(args) + + def __call__( + self, + inputs: mx.array, + cache=None, + ): + h = self.mixer(self.norm(inputs), cache=cache) + r = inputs + h + return r + + +class Mamba2(nn.Module): + def __init__(self, args: ModelArgs): + super().__init__() + self.args = args + self.embeddings = nn.Embedding(args.vocab_size, args.hidden_size) + self.layers = [Mamba2Block(args) for idx in range(args.num_hidden_layers)] + self.norm_f = nn.RMSNorm(args.hidden_size, eps=args.layer_norm_epsilon) + + def __call__( + self, + inputs: mx.array, + cache=None + ): + hidden_states = self.embeddings(inputs) + + if cache is None: + cache = Mamba2Cache(len(self.layers)) + + for i, layer in enumerate(self.layers): + hidden_states = layer(hidden_states, cache[i]) + + hidden_states = self.norm_f(hidden_states) + return hidden_states + + +class Model(nn.Module): + def __init__(self, args: ModelArgs): + super().__init__() + self.args = args + self.model_type = args.model_type + self.backbone = Mamba2(args) + if not args.tie_word_embeddings: + self.lm_head = nn.Linear(args.hidden_size, args.vocab_size, bias=False) + + def __call__( + self, + inputs: mx.array, + cache=None + ): + B, T = inputs.shape + + x = self.backbone(inputs, cache) + + if self.args.tie_word_embeddings: + logits = self.backbone.embeddings.as_linear(x) + else: + logits = self.lm_head(x) + return logits + + def sanitize(self, weights): + for k, v in weights.items(): + if "conv1d.weight" in k and v.ndim == 3: + weights[k] = v.moveaxis(2, 1) + return weights + + def make_cache(self, batch_size: int = 1): + return Mamba2Cache(len(self.backbone.layers)) + + @property + def layers(self): + return self.backbone.layers diff --git a/llms/mlx_lm/models/mamba2.py b/llms/mlx_lm/models/mamba2.py index f74ae826..8b0ee59e 100644 --- a/llms/mlx_lm/models/mamba2.py +++ b/llms/mlx_lm/models/mamba2.py @@ -2,7 +2,7 @@ import math from dataclasses import dataclass, field -from typing import Optional, Tuple, Union +from typing import Tuple, Union import mlx.core as mx import mlx.nn as nn @@ -20,15 +20,11 @@ class ModelArgs(BaseModelArgs): state_size: int = 128 num_hidden_layers: int = 64 layer_norm_epsilon: float = 1e-5 - pad_token_id: int = 1 - bos_token_id: int = 0 - eos_token_id: int = 2 expand: int = 2 conv_kernel: int = 4 n_groups: int = 8 use_bias: bool = False use_conv_bias: bool = True - hidden_act: str = "silu" initializer_range: float = 0.1 residual_in_fp32: bool = True time_step_rank: Union[int, str] = "auto" @@ -52,14 +48,18 @@ class ModelArgs(BaseModelArgs): class Mamba2Cache: - def __init__(self, num_layers): - self.cache = [[None, None] for _ in range(num_layers)] + def __init__(self): + self.cache = [None, None] + + def __setitem__(self, idx, value): + self.cache[idx] = value def __getitem__(self, idx): return self.cache[idx] - def __setitem__(self, idx, value): - self.cache[idx] = value + @property + def state(self): + return self.cache class MambaRMSNormGated(nn.Module): @@ -74,67 +74,54 @@ class MambaRMSNormGated(nn.Module): variance = mx.mean(hidden_states ** 2, axis=-1, keepdims=True) hidden_states = hidden_states * mx.rsqrt(variance + self.variance_epsilon) return self.weight * hidden_states - class DepthWiseConv1d(nn.Module): def __init__(self, in_channels, out_channels, kernel_size, bias=True, groups=None, padding=0): super().__init__() - assert in_channels == out_channels, "For depthwise conv, in_channels must equal out_channels" - self.channels = in_channels + self.in_channels = in_channels + self.out_channels = out_channels self.kernel_size = kernel_size self.padding = padding - - # For depthwise conv, we use groups equal to the number of channels - self.groups = self.channels if groups is None else groups - assert self.groups == self.channels, "For depthwise conv, groups must equal the number of channels" + self.groups = groups if groups is not None else in_channels - # Weight shape: (channels, 1, kernel_size) for depthwise conv - self.weight = mx.random.normal((self.channels, 1, kernel_size)) - self.bias = mx.zeros((self.channels,)) if bias else None + # Ensure in_channels and out_channels are the same for depthwise conv + assert in_channels == out_channels, "In and out channels must be the same for depthwise convolution" + # Ensure groups is equal to in_channels for depthwise conv + assert self.groups == in_channels, "Groups must be equal to in_channels for depthwise convolution" + + # Initialize weight with shape (out_channels, kernel_size, 1) + self.weight = mx.random.normal((out_channels, kernel_size, 1)) + self.bias = mx.zeros((out_channels,)) if bias else None def __call__(self, x, cache=None): B, L, C = x.shape - K = self.kernel_size + _, K, _ = self.weight.shape if cache is not None: x = mx.concatenate([cache, x], axis=1) else: - x = mx.pad(x, [(0, 0), (self.padding, 0), (0, 0)]) + x = mx.pad(x, [(0, 0), (K - 1, 0), (0, 0)]) - # Reshape for depthwise convolution - x = x.transpose(0, 2, 1) # (B, C, L) - - # Perform depthwise convolution - y = mx.conv(x, self.weight, groups=self.groups) - - # Reshape back - y = y.transpose(0, 2, 1) # (B, L, C) + y = mx.conv_general(x, self.weight, groups=self.groups) if self.bias is not None: y = y + self.bias - return y, x.transpose(0, 2, 1)[:, -K:, :] + return y, x[:, -K + 1 :, :] class Mamba2Mixer(nn.Module): def __init__(self, args: ModelArgs): super().__init__() self.args = args - self.hidden_size = args.hidden_size self.intermediate_size = args.intermediate_size + self.time_step_rank = args.time_step_rank self.conv_kernel_size = args.conv_kernel + self.hidden_size = args.hidden_size self.state_size = args.state_size self.num_heads = args.num_heads self.head_dim = args.head_dim self.n_groups = args.n_groups - self.time_step_rank = args.time_step_rank - - projection_size = self.intermediate_size + self.intermediate_size + 2 * self.n_groups * self.state_size + self.num_heads - self.in_proj = nn.Linear( - self.hidden_size, - projection_size, - bias=args.use_bias - ) self.conv_dim = self.intermediate_size + 2 * self.n_groups * self.state_size self.conv1d = DepthWiseConv1d( @@ -143,32 +130,74 @@ class Mamba2Mixer(nn.Module): bias=args.use_conv_bias, kernel_size=args.conv_kernel, groups=self.conv_dim, - padding=args.conv_kernel - 1, + padding=args.conv_kernel - 1 + ) + + projection_size = self.intermediate_size + self.conv_dim + self.num_heads + self.in_proj = nn.Linear( + self.hidden_size, + projection_size, + bias=args.use_bias ) self.act = nn.SiLU() self.dt_bias = mx.ones((self.num_heads,)) - self.A_log = mx.log(mx.arange(1, self.num_heads + 1, dtype=mx.float32)) + self.A_log = mx.log(mx.arange(1, self.num_heads + 1)) self.D = mx.ones((self.num_heads,)) - self.out_proj = nn.Linear(self.intermediate_size, self.hidden_size, bias=args.use_bias) self.norm = MambaRMSNormGated(self.intermediate_size, eps=args.layer_norm_epsilon) + self.out_proj = nn.Linear(self.intermediate_size, self.hidden_size, bias=args.use_bias) + + # def ssm_step(self, x, state=None): + # A = -mx.exp(self.A_log) + # D = self.D + # deltaBC = self.x_proj(x) + # delta, B, C = mx.split( + # deltaBC, + # indices_or_sections=[ + # self.time_step_rank, + # self.time_step_rank + self.ssm_state_size, + # ], + # axis=-1, + # ) + # delta = nn.softplus(self.dt_proj(delta)) + # new_state = mx.expand_dims(delta * x, -1) * mx.expand_dims(B, 1) + # if state is not None: + # new_state += state * mx.exp(mx.expand_dims(delta, -1) * A) + # y = (new_state @ mx.expand_dims(C, -1)).squeeze(2) + # y = y + D * x + # return y, new_state + def ssm_step(self, x, dt, state): - A = -mx.exp(self.A_log) - D = self.D + B, L, C = x.shape + print(f"x shape: {x.shape}") + projected_states = self.in_proj(x) + print(f"deltaBC shape: {projected_states.shape}") - deltaBC = self.in_proj(x) - gate, conv_state, time_step = mx.split( - deltaBC, - [self.intermediate_size, self.intermediate_size + 2 * self.n_groups * self.state_size], - axis=-1 - ) + d_mlp = (projected_states.shape[-1] - 2 * self.intermediate_size - 2 * self.n_groups * self.state_size - self.num_heads) // 2 - conv_state = conv_state.transpose(0, 2, 1) + gate = projected_states[:, :, 2*d_mlp:2*d_mlp+self.intermediate_size] + conv_state = projected_states[:, :, 2*d_mlp+self.intermediate_size:2*d_mlp+self.intermediate_size+self.conv_dim] + time_step = projected_states[:, :, -self.num_heads:] + + print(f"conv_state shape before reshape: {conv_state.shape}") + print(f"self.conv_dim: {self.conv_dim}") + + # Reshape and handle the case where L=1 + conv_state = conv_state.reshape(B, self.conv_dim, L) + if L == 1: + # If sequence length is 1, we need to pad to apply convolution + conv_state = mx.pad(conv_state, ((0, 0), (0, 0), (0, self.conv_kernel_size - 1))) + conv_out = self.conv1d(conv_state) + + # If we padded, we need to remove the padding + if L == 1: + conv_out = conv_out[:, :, :L] + + # Reshape back to (B, L, C) conv_out = conv_out.transpose(0, 2, 1) - conv_out = self.act(conv_out) x_and_conv_out, B, C = mx.split( conv_out, @@ -187,58 +216,47 @@ class Mamba2Mixer(nn.Module): new_state = state * dA + x_and_conv_out[:, :, None, None] * dB y = mx.sum(new_state * C, axis=-1) - y = y + D[None, :, None] * x_and_conv_out + y = y + C[None, :, None] * x_and_conv_out y = self.norm(y.reshape(-1, self.intermediate_size), gate) output = self.out_proj(y) return output, new_state - def __call__( - self, - x: mx.array, - cache = None - ): - B, L, _ = x.shape + def __call__(self, x, cache): + B, T, D = x.shape + if cache is None: + cache = [None, None] - if cache[0] is not None: # Using cached state - conv_state, ssm_state = cache - x = x[:, -1:] - output, new_ssm_state = self.ssm_step(x, None, ssm_state) - cache[1] = new_ssm_state # Update SSM state in cache - else: - conv_state, ssm_state = None, None - outputs = [] - for t in range(L): - x = x[:, t:t+1] - output, ssm_state = self.ssm_step(x, None, ssm_state) - outputs.append(output) - output = mx.concatenate(outputs, axis=1) - cache[1] = ssm_state # Store final SSM state in cache + outputs = [] + for t in range(T): + xt = x[:, t, :] + xz = self.in_proj(xt) + x_t, z_t = xz.split(indices_or_sections=2, axis=1) - # Update conv state in cache - new_conv_state = x[:, -self.conv_kernel_size:] - cache[0] = new_conv_state + if x_t.shape[-1] != self.conv_dim: + raise ValueError(f"Expected conv input dim {self.conv_dim}, got {x_t.shape[-1]}") + conv_out, cache[0] = self.conv1d(mx.expand_dims(x_t, 1), cache[0]) + x_t = conv_out.squeeze(1) + x_t = nn.silu(x_t) + y_t, cache[1] = self.ssm_step(x_t, cache[1]) + z_t = nn.silu(z_t) + output_t = y_t * z_t + output_t = self.out_proj(output_t) + outputs.append(output_t) + output = mx.stack(outputs, axis=1) return output class Mamba2Block(nn.Module): def __init__(self, args: ModelArgs): super().__init__() - self.args = args - self.residual_in_fp32 = args.residual_in_fp32 - self.norm = nn.RMSNorm(args.hidden_size, eps=args.layer_norm_epsilon) self.mixer = Mamba2Mixer(args) + self.norm = nn.RMSNorm(args.hidden_size) - def __call__( - self, - inputs: mx.array, - cache=None, - ): - h = self.mixer(self.norm(inputs), cache_params=cache) - r = inputs + h - return r + def __call__(self, x: mx.array, cache): + return self.mixer(self.norm(x), cache) + x class Mamba2(nn.Module): @@ -275,11 +293,7 @@ class Model(nn.Module): if not args.tie_word_embeddings: self.lm_head = nn.Linear(args.hidden_size, args.vocab_size, bias=False) - def __call__( - self, - inputs: mx.array, - cache=None - ): + def __call__(self, inputs: mx.array, cache=None): B, T = inputs.shape x = self.backbone(inputs, cache) @@ -288,17 +302,18 @@ class Model(nn.Module): logits = self.backbone.embeddings.as_linear(x) else: logits = self.lm_head(x) + return logits - - def sanitize_mabey(self, weights): + + def sanitize(self, weights): for k, v in weights.items(): if "conv1d.weight" in k and v.ndim == 3: weights[k] = v.moveaxis(2, 1) return weights - + def make_cache(self, batch_size: int = 1): - return Mamba2Cache(len(self.backbone.layers)) - + return [Mamba2Cache() for _ in range(len(self.layers))] + @property def layers(self): return self.backbone.layers