add support for cohere2

llms/mlx_lm/models/cohere2.py

@@ -0,0 +1,207 @@
+# Copyright © 2023-2024 Apple Inc.
+
+from dataclasses import dataclass
+from typing import Optional, Tuple
+
+import mlx.core as mx
+import mlx.nn as nn
+
+from .base import BaseModelArgs, create_attention_mask
+
+
+@dataclass
+class ModelArgs(BaseModelArgs):
+    model_type: str
+    hidden_size: int = 4096
+    head_dim: int = 128
+    num_hidden_layers: int = 32
+    intermediate_size: int = 14336
+    num_attention_heads: int = 32
+    num_key_value_heads: int = 8
+    rope_theta: float = 50000.0
+    vocab_size: int = 256000
+    layer_norm_eps: float = 1e-05
+    logit_scale: float = 0.0625
+    attention_bias: bool = False
+    layer_norm_bias: bool = False
+    sliding_window: int = 4096
+    sliding_window_pattern: int = 4
+
+
+class LayerNorm2D(nn.Module):
+
+    def __init__(self, d1, d2, eps):
+        super().__init__()
+        self.weight = mx.zeros((d1, d2))
+        self.eps = eps
+
+    def __call__(self, x):
+        return self.weight * mx.fast.layer_norm(x, None, None, self.eps)
+
+
+class Attention(nn.Module):
+    def __init__(self, args: ModelArgs, layer_idx: int):
+        super().__init__()
+        self.args = args
+        self.layer_idx = layer_idx
+
+        dim = args.hidden_size
+        self.n_heads = n_heads = args.num_attention_heads
+        self.n_kv_heads = n_kv_heads = args.num_key_value_heads
+
+        head_dim = args.hidden_size // args.num_attention_heads
+        self.scale = head_dim**-0.5
+
+        attetion_bias = args.attention_bias
+
+        self.q_proj = nn.Linear(dim, n_heads * head_dim, bias=attetion_bias)
+        self.k_proj = nn.Linear(dim, n_kv_heads * head_dim, bias=attetion_bias)
+        self.v_proj = nn.Linear(dim, n_kv_heads * head_dim, bias=attetion_bias)
+        self.o_proj = nn.Linear(n_heads * head_dim, dim, bias=attetion_bias)
+
+        self.rope = nn.RoPE(head_dim, traditional=True, base=args.rope_theta)
+
+        self.sliding_window = (
+            args.sliding_window
+            if (layer_idx + 1) % args.sliding_window_pattern != 0
+            else None
+        )
+
+    def __call__(
+        self,
+        x: mx.array,
+        mask: Optional[mx.array] = None,
+        cache: Optional[Tuple[mx.array, mx.array]] = None,
+    ) -> mx.array:
+        B, L, D = x.shape
+
+        queries, keys, values = self.q_proj(x), self.k_proj(x), self.v_proj(x)
+
+        queries = queries.reshape(B, L, self.n_heads, -1)
+        keys = keys.reshape(B, L, self.n_kv_heads, -1)
+
+        queries = queries.transpose(0, 2, 1, 3)
+        keys = keys.transpose(0, 2, 1, 3)
+        values = values.reshape(B, L, self.n_kv_heads, -1).transpose(0, 2, 1, 3)
+
+        if cache is not None:
+            queries = self.rope(queries, offset=cache.offset)
+            keys = self.rope(keys, offset=cache.offset)
+            keys, values = cache.update_and_fetch(keys, values)
+        else:
+            queries = self.rope(queries)
+            keys = self.rope(keys)
+
+            # sliding window attention
+            if self.sliding_window is not None:
+                keys = keys[:, : -self.sliding_window :, :]
+                values = values[:, : -self.sliding_window :, :]
+                if mask is not None:
+                    mask = mask[:, : -self.sliding_window, :]
+
+        output = mx.fast.scaled_dot_product_attention(
+            queries, keys, values, scale=self.scale, mask=mask
+        )
+
+        output = output.transpose(0, 2, 1, 3).reshape(B, L, -1)
+        return self.o_proj(output)
+
+
+class MLP(nn.Module):
+    def __init__(self, dim, hidden_dim):
+        super().__init__()
+        self.gate_proj = nn.Linear(dim, hidden_dim, bias=False)
+        self.up_proj = nn.Linear(dim, hidden_dim, bias=False)
+        self.down_proj = nn.Linear(hidden_dim, dim, bias=False)
+
+    def __call__(self, x):
+        return self.down_proj(nn.gelu(self.gate_proj(x)) * self.up_proj(x))
+
+
+class TransformerBlock(nn.Module):
+    def __init__(self, args: ModelArgs, layer_idx: int):
+        super().__init__()
+        self.hidden_size = args.hidden_size
+        self.n_heads = args.num_attention_heads
+
+        self.self_attn = Attention(args, layer_idx)
+        self.mlp = MLP(args.hidden_size, args.intermediate_size)
+        self.input_layernorm = nn.LayerNorm(
+            args.hidden_size, eps=args.layer_norm_eps, bias=args.layer_norm_bias
+        )
+        self.args = args
+
+    def __call__(
+        self,
+        x: mx.array,
+        mask: Optional[mx.array] = None,
+        cache: Optional[Tuple[mx.array, mx.array]] = None,
+    ) -> mx.array:
+        h = self.input_layernorm(x)
+        attn_h = self.self_attn(h, mask, cache)
+        ff_h = self.mlp(h)
+        return attn_h + ff_h + x
+
+
+class CohereModel(nn.Module):
+    def __init__(self, args: ModelArgs):
+        super().__init__()
+        self.args = args
+        self.vocab_size = args.vocab_size
+        self.num_hidden_layers = args.num_hidden_layers
+        assert self.vocab_size > 0
+        self.embed_tokens = nn.Embedding(args.vocab_size, args.hidden_size)
+        self.layers = [
+            TransformerBlock(args=args, layer_idx=i)
+            for i in range(args.num_hidden_layers)
+        ]
+        self.norm = nn.LayerNorm(
+            args.hidden_size, eps=args.layer_norm_eps, bias=args.layer_norm_bias
+        )
+
+    def __call__(
+        self,
+        inputs: mx.array,
+        cache=None,
+    ):
+        h = self.embed_tokens(inputs)
+
+        mask = create_attention_mask(h, cache)
+
+        if cache is None:
+            cache = [None] * len(self.layers)
+
+        for layer, c in zip(self.layers, cache):
+            h = layer(h, mask, c)
+
+        return self.norm(h)
+
+
+class Model(nn.Module):
+    def __init__(self, args: ModelArgs):
+        super().__init__()
+        self.model_type = args.model_type
+        self.model = CohereModel(args)
+        self.args = args
+
+    def __call__(
+        self,
+        inputs: mx.array,
+        cache=None,
+    ):
+        out = self.model(inputs, cache)
+        out = self.model.embed_tokens.as_linear(out)
+        out = out * self.model.args.logit_scale
+        return out
+
+    @property
+    def layers(self):
+        return self.model.layers
+
+    @property
+    def head_dim(self):
+        return self.args.hidden_size // self.args.num_attention_heads
+
+    @property
+    def n_kv_heads(self):
+        return self.args.num_key_value_heads