mistral_v0_2.model.attention#

mistral_v0_2.model.attention.convert_attention_params(self_attn)[source]#

Converts the attention parameters from MistralAttention HuggingFace format with PyTorch tensor to jax.Array.

Parameters:

self_attn (MistralAttention) – The attention parameters in MistralAttention.

Returns:

The attention parameters converted into the AttentionParams with JAX.

Return type:

AttentionParams

mistral_v0_2.model.attention.forward_attention(params, seq, qk_mask, rotary_values, kv_cache_cur, kv_cache_pre)[source]#

Performs the forward pass of the attention mechanism using.

This function executes the attention mechanism on the input sequence seq using the provided attention parameters.

Parameters:
  • params (AttentionParams) – The attention parameters.

  • seq (Array) – The input sequences on which attention is to be applied.

  • qk_mask (Array) – The qk mask for the attention mechanism, determining which parts of the sequence are allowed to attend to each other.

  • rotary_values (RotaryValues) – Rotary positional embeddings values.

  • kv_cache_cur (KVCache) – The current KVCache.

  • kv_cache_pre (KVCache) – The previous KVCache.

Returns:

A tuple containing the output sequence after applying attention, and the updated current and previous KVCache.

Return type:

tuple[Array, KVCache, KVCache]

mistral_v0_2.model.attention.shard_attention_params(params)[source]#

Shard the attention parameters for distributed computing.

Parameters:

params (AttentionParams) – The attention parameters.

Returns:

The attention parameters modified with tensor parallelism, allowing for distributed computation across multiple devices.

Return type:

AttentionParams

mistral_v0_2.model.attention.test_forward_attention(model)[source]#

Tests the forward attention mechanism.

This function is designed to validate the functionality and correctness of the attention mechanism with JAX.

Parameters:

model (MistralForCausalLM) – PyTorch Mistral model to compare with this implementation.

Return type:

None

Returns:

None.