huggingface/text-generation-inference
5
0
Fork 0
mirror of https://github.com/huggingface/text-generation-inference.git synced 2025-09-10 20:04:52 +00:00
Code Issues Packages Projects Releases Wiki Activity

nit

Browse Source
This commit is contained in:
Felix Marty 2023-07-19 16:31:46 +00:00
parent 5a1512c025
commit 5882768682
1 changed files with 19 additions and 15 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

22
server/text_generation_server/models/custom_modeling/flash_llama_modeling.py
View File

@ -148,23 +148,27 @@ class LlamaRMSNorm(nn.Module):
def _load_gqa(config, prefix: str, weights): def _load_gqa(config, prefix: str, weights):
w = [
weights.get_sharded(f"{prefix}.q_proj.weight", dim=0),
weights.get_sharded(f"{prefix}.k_proj.weight", dim=0),
weights.get_sharded(f"{prefix}.v_proj.weight", dim=0),
]
weight = torch.cat(w, dim=0)
weight = weight.to(dtype=weights.dtype).to(device=weights.device)
bias = None
assert config.hidden_size % config.num_attention_heads == 0 assert config.hidden_size % config.num_attention_heads == 0
head_size = config.hidden_size // config.num_attention_heads
assert config.num_attention_heads % weights.process_group.size() == 0 assert config.num_attention_heads % weights.process_group.size() == 0
weight = weights.get_multi_weights_col(
prefixes=[f"{prefix}.q_proj", f"{prefix}.k_proj", f"{prefix}.v_proj"],
quantize=config.quantize,
dim=0
)
if config.quantize != "gptq":
weight = weight.to(dtype=weights.dtype).to(device=weights.device)
head_size = config.hidden_size // config.num_attention_heads
num_heads = config.num_attention_heads // weights.process_group.size() num_heads = config.num_attention_heads // weights.process_group.size()
num_key_value_heads = config.num_key_value_heads // weights.process_group.size() num_key_value_heads = config.num_key_value_heads // weights.process_group.size()
assert list(weight.shape) == [ assert list(weight.shape) == [
(num_heads + 2 * num_key_value_heads) * head_size, (num_heads + 2 * num_key_value_heads) * head_size,
config.hidden_size, config.hidden_size,
], f"{list(weight.shape)} != {[(num_heads + 2 * config.num_key_value_heads) * head_size, config.hidden_size]}" ], f"{list(weight.shape)} != {[(num_heads + 2 * config.num_key_value_heads) * head_size, config.hidden_size]}"
bias = None
return TensorParallelColumnLinear(get_linear(weight, bias, config.quantize)) return TensorParallelColumnLinear(get_linear(weight, bias, config.quantize))