| /aosp_15_r20/external/executorch/examples/mediatek/model_export_scripts/ |
| H A D | llama.py | 165 num_chunks, num_key_value_heads, num_blocks_per_chunk, head_dim, max_cache_size argument
232 head_dim, argument
|
| /aosp_15_r20/external/pytorch/benchmarks/gpt_fast/ |
| H A D | model.py | 26 head_dim: int = 64 variable in ModelArgs
93 self, max_batch_size, max_seq_length, n_heads, head_dim, dtype=torch.bfloat16 argument
|
| H A D | mixtral_moe_model.py | 26 head_dim: int = 64 variable in ModelArgs
72 self, max_batch_size, max_seq_length, n_heads, head_dim, dtype=torch.bfloat16 argument
|
| /aosp_15_r20/external/pytorch/test/ |
| H A D | test_flop_counter.py | 323 head_dim, argument
480 head_dim, argument
|
| H A D | test_transformers.py | 1703 def _get_block_size_n(device, head_dim, is_dropout, is_causal): argument
1811 head_dim, argument
1894 head_dim, argument
3326 def rand_nt(sequence_list, num_heads, head_dim): argument
|
| /aosp_15_r20/external/executorch/backends/vulkan/test/op_tests/ |
| H A D | sdpa_test.cpp | 222 const int head_dim = embedding_dim / num_heads; in test_reference_sdpa() local
292 const int head_dim = embedding_dim / num_heads; in test_vulkan_sdpa() local
|
| H A D | rotary_embedding_test.cpp | 83 const int head_dim = dim / n_heads; in test_reference() local
|
| /aosp_15_r20/external/executorch/examples/mediatek/models/llm_models/weights/Llama-3.2-1B-Instruct/ |
| H A D | config.json | 7 "head_dim": 64, number
|
| /aosp_15_r20/external/executorch/examples/mediatek/models/llm_models/weights/Llama-3.2-3B-Instruct/ |
| H A D | config.json | 7 "head_dim": 128, number
|
| /aosp_15_r20/external/pytorch/benchmarks/transformer/ |
| H A D | sdpa.py | 40 def head_dim(self) -> int: member in ExperimentConfig
|
| H A D | attention_bias_benchmarks.py | 39 def head_dim(self) -> int: member in ExperimentConfig
|
| /aosp_15_r20/external/executorch/examples/models/llama/source_transformation/ |
| H A D | quantized_kv_cache.py | 36 head_dim, argument
|
| H A D | apply_spin_quant_r1_r2.py | 68 def rotate_ov_proj(layer, head_dim, R2=None): argument
|
| /aosp_15_r20/external/pytorch/torch/csrc/api/include/torch/nn/modules/ |
| H A D | activation.h | 864 int64_t head_dim; variable
|
| /aosp_15_r20/external/executorch/extension/llm/custom_ops/ |
| H A D | test_sdpa_with_kv_cache.py | 407 head_dim, argument
|
| /aosp_15_r20/external/pytorch/torch/csrc/api/include/torch/nn/functional/ |
| H A D | activation.h | 677 const auto head_dim = embed_dim / num_heads; variable
|
| /aosp_15_r20/external/pytorch/aten/src/ATen/native/nested/cuda/ |
| H A D | NestedTensorTransformerUtils.cpp | 174 const int64_t head_dim, in view_as_dense()
|
| /aosp_15_r20/external/pytorch/aten/src/ATen/native/transformers/cuda/ |
| H A D | sdp_utils.cpp | 316 const auto head_dim = params.query.sym_size(3); in check_cudnn_tensor_shapes() local
|
| /aosp_15_r20/external/pytorch/aten/src/ATen/native/transformers/cuda/mem_eff_attention/ |
| H A D | kernel_forward.h | 158 int32_t head_dim = 0; member
|
| H A D | kernel_backward.h | 662 int32_t head_dim = -1; member
|
| /aosp_15_r20/external/pytorch/test/inductor/ |
| H A D | test_aot_inductor.py | 1444 head_dim, argument
|