x

research/long-context-llama/H2O/cache_utils.py

@@ -137,7 +137,6 @@ class DynamicCache(Cache):
             self._seen_tokens += key_states.shape[-2]
 
         # Update the cache
-        print(len(self.key_cache), layer_idx)
         if len(self.key_cache) <= layer_idx:
             self.key_cache.append(key_states)
             self.value_cache.append(value_states)
@@ -340,6 +339,163 @@ class SinkCache(Cache):
             self.value_cache[layer_idx] = self.value_cache[layer_idx].index_select(0, beam_idx.to(device))
 
 
+class HHCache(Cache):
+    """
+    A cache that as described in the [Attention Sinks paper](https://arxiv.org/abs/2309.17453). It allows the model to
+    generate beyond the length of its context window, without losing fluency in the conversation. As it discards past
+    tokens, the model will lose the ability to generate tokens that depend on the context that was discarded.
+
+    It stores the Key and Value states as a list of tensors, one for each layer. The expected shape for each tensor is
+    `[batch_size, num_heads, seq_len, head_dim]`.
+
+    Parameters:
+        window_length (`int`):
+            The length of the context window.
+        num_sink_tokens (`int`):
+            The number of sink tokens. See the original paper for more information.
+    """
+
+    def __init__(self, window_length: int, num_sink_tokens: int) -> None:
+        self.key_cache: List[torch.Tensor] = []
+        self.value_cache: List[torch.Tensor] = []
+        self.window_length = window_length
+        self.num_sink_tokens = num_sink_tokens
+        self.cos_sin_cache = {}
+        self._seen_tokens = 0  # Used in `generate` to keep tally of how many tokens the cache has seen
+
+    @staticmethod
+    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_key_rotary_pos_emb(
+        self, key_states: torch.Tensor, cos: torch.Tensor, sin: torch.Tensor
+    ) -> torch.Tensor:
+        rotated_key_states = (key_states * cos) + (self._rotate_half(key_states) * sin)
+        return rotated_key_states
+
+    def _get_rerotation_cos_sin(
+        self, key_states: torch.Tensor, cos: torch.Tensor, sin: torch.Tensor
+    ) -> Tuple[torch.Tensor, torch.Tensor]:
+        if key_states.shape[-2] not in self.cos_sin_cache:
+            # Upcast to float32 temporarily for better accuracy
+            cos = cos.to(torch.float32)
+            sin = sin.to(torch.float32)
+
+            # Compute the cos and sin required for back- and forward-rotating to one position earlier in the sequence
+            original_cos = cos[self.num_sink_tokens + key_states.shape[-2] :]
+            shifted_cos = cos[self.num_sink_tokens : -key_states.shape[-2]]
+            original_sin = sin[self.num_sink_tokens + key_states.shape[-2] :]
+            shifted_sin = sin[self.num_sink_tokens : -key_states.shape[-2]]
+            rerotation_cos = original_cos * shifted_cos + original_sin * shifted_sin
+            rerotation_sin = -original_sin * shifted_cos + original_cos * shifted_sin
+
+            self.cos_sin_cache[key_states.shape[-2]] = (
+                rerotation_cos.to(key_states.dtype).unsqueeze(0),
+                rerotation_sin.to(key_states.dtype).unsqueeze(0),
+            )
+        return self.cos_sin_cache[key_states.shape[-2]]
+
+    def get_seq_length(self, layer_idx: Optional[int] = 0) -> int:
+        """Returns the sequence length of the cached states. A layer index can be optionally passed."""
+        # Workaround to make 'key_states.shape[-2] + past_key_value.get_seq_length(self.layer_idx)' <= window_length
+        if len(self.key_cache) <= layer_idx:
+            return 0
+        return self.key_cache[layer_idx].shape[-2]
+
+    def get_max_length(self) -> Optional[int]:
+        """Returns the maximum sequence length of the cached states."""
+        return self.window_length
+
+    def update(
+        self,
+        key_states: torch.Tensor,
+        value_states: torch.Tensor,
+        layer_idx: int,
+        cache_kwargs: Optional[Dict[str, Any]] = None,
+    ) -> Tuple[torch.Tensor, torch.Tensor]:
+        """
+        Updates the cache with the new `key_states` and `value_states` for the layer `layer_idx`.
+
+        Parameters:
+            key_states (`torch.Tensor`):
+                The new key states to cache.
+            value_states (`torch.Tensor`):
+                The new value states to cache.
+            layer_idx (`int`):
+                The index of the layer to cache the states for.
+            cache_kwargs (`Dict[str, Any]`, `optional`):
+                Additional arguments for the cache subclass. The following arguments can be used in `SinkCache`: `sin`,
+                `cos` and `partial_rotation_size`. These arguments are used with models using RoPE, to recompute the
+                rotation as the tokens are shifted.
+
+        Return:
+            A tuple containing the updated key and value states.
+        """
+        # Optional kwargs for `SinkCache` -- needed on models using RoPE. `partial_rotation_size` is used on models
+        # with partially rotated position embeddings, like Phi or Persimmon.
+        sin = cache_kwargs.get("sin")
+        cos = cache_kwargs.get("cos")
+        partial_rotation_size = cache_kwargs.get("partial_rotation_size")
+        using_rope = cos is not None and sin is not None
+
+        # Update the number of seen tokens
+        if layer_idx == 0:
+            self._seen_tokens += key_states.shape[-2]
+
+        # [bsz, num_heads, seq_len, head_dim]
+        if len(self.key_cache) <= layer_idx:
+            # Empty cache
+            self.key_cache.append(key_states)
+            self.value_cache.append(value_states)
+
+        elif key_states.shape[-2] + self.get_seq_length(layer_idx) < self.window_length:
+            # Growing cache
+            self.key_cache[layer_idx] = torch.cat([self.key_cache[layer_idx], key_states], dim=-2)
+            self.value_cache[layer_idx] = torch.cat([self.value_cache[layer_idx], value_states], dim=-2)
+
+        else:
+            # Shifting cache
+            keys_to_keep = self.key_cache[layer_idx][
+                :, :, -self.window_length + self.num_sink_tokens + key_states.shape[-2] :
+            ]
+
+            # On RoPE models, we need to recompute the Key rotation as the tokens are shifted
+            if using_rope:
+                rerotation_cos, rerotation_sin = self._get_rerotation_cos_sin(
+                    key_states, cos[: self.window_length], sin[: self.window_length]
+                )
+                if partial_rotation_size is not None:
+                    keys_to_keep, keys_pass = (
+                        keys_to_keep[..., :partial_rotation_size],
+                        keys_to_keep[..., partial_rotation_size:],
+                    )
+                keys_to_keep = self._apply_key_rotary_pos_emb(keys_to_keep, rerotation_cos, rerotation_sin)
+                if partial_rotation_size is not None:
+                    keys_to_keep = torch.cat((keys_to_keep, keys_pass), dim=-1)
+
+            # Concatenate sink tokens, shifted & rotated tokens (if needed), and new tokens
+            sink_keys = self.key_cache[layer_idx][:, :, : self.num_sink_tokens]
+            self.key_cache[layer_idx] = torch.cat([sink_keys, keys_to_keep, key_states], dim=-2)
+
+            sink_values = self.value_cache[layer_idx][:, :, : self.num_sink_tokens]
+            values_to_keep = self.value_cache[layer_idx][
+                :, :, -self.window_length + self.num_sink_tokens + value_states.shape[-2] :
+            ]
+            self.value_cache[layer_idx] = torch.cat([sink_values, values_to_keep, value_states], dim=-2)
+
+        return self.key_cache[layer_idx], self.value_cache[layer_idx]
+
+    def reorder_cache(self, beam_idx: torch.LongTensor):
+        """Reorders the cache for beam search, given the selected beam indices."""
+        for layer_idx in range(len(self.key_cache)):
+            device = self.key_cache[layer_idx].device
+            self.key_cache[layer_idx] = self.key_cache[layer_idx].index_select(0, beam_idx.to(device))
+            device = self.value_cache[layer_idx].device
+            self.value_cache[layer_idx] = self.value_cache[layer_idx].index_select(0, beam_idx.to(device))
+
+
 class StaticCache(Cache):
     """
     Static Cache class to be used with `torch.compile(model)`.

research/long-context-llama/H2O/utils_llama.py

@@ -538,4 +538,4 @@ class H2OLlamaForCausalLM(LlamaForCausalLM):
         super().__init__(config)
         num_layers = len(self.model.layers)
         for layer_idx in range(num_layers):
-            self.model.layers[layer_idx].self_attn = H2OLlamaAttention(config)
+            self.model.layers[layer_idx].self_attn = H2OLlamaAttention(config, layer_idx)