apps/FastDeploy
1
0
Fork 0
mirror of https://github.com/PaddlePaddle/FastDeploy.git synced 2025-10-06 00:57:33 +08:00
Code Issues Actions 2 Packages Projects Releases Wiki Activity

[BugFix]Fix load kv cache quant scale (#4077)

Browse Source 
* fix kv cache

* fix kv_cache

* fix kv cache
This commit is contained in:
YuanRisheng
2025-09-12 17:44:03 +08:00
committed by GitHub
parent c86b3357ce
commit 88ea565aba
2 changed files with 10 additions and 20 deletions
Download Patch File Download Diff File Expand all files Collapse all files

15
fastdeploy/model_executor/layers/quantization/kv_cache.py
View File

@@ -63,6 +63,7 @@ class KvCacheQuantConfig(QuantConfigBase):
if self.quant_type == KvCacheQuantzationTypes.INT8 or self.quant_type == KvCacheQuantzationTypes.INT8_ZP: if self.quant_type == KvCacheQuantzationTypes.INT8 or self.quant_type == KvCacheQuantzationTypes.INT8_ZP:
self.max_bound = 127.0 self.max_bound = 127.0
self.is_channel_wise = True
elif ( elif (
self.quant_type == KvCacheQuantzationTypes.FP8 self.quant_type == KvCacheQuantzationTypes.FP8
or self.quant_type == KvCacheQuantzationTypes.FP8_ZP or self.quant_type == KvCacheQuantzationTypes.FP8_ZP
@@ -125,18 +126,6 @@ class KVCacheMethodBase(QuantMethodBase):
load_scale load_scale
""" """
if self.cache_quant_config.is_channel_wise:
cache_k_scale_tensor = (
get_tensor(state_dict.pop(self.cache_k_scale_name))
.cast(paddle.get_default_dtype())
.reshape_([-1, layer.head_dim])
)
cache_v_scale_tensor = (
get_tensor(state_dict.pop(self.cache_v_scale_name))
.cast(paddle.get_default_dtype())
.reshape_([-1, layer.head_dim])
)
else:
cache_k_scale_tensor = ( cache_k_scale_tensor = (
get_tensor(state_dict.pop(self.cache_k_scale_name)).cast(paddle.get_default_dtype()).reshape_([-1]) get_tensor(state_dict.pop(self.cache_k_scale_name)).cast(paddle.get_default_dtype()).reshape_([-1])
) )
@@ -185,7 +174,7 @@ class KVCacheMethodBase(QuantMethodBase):
scale_shape = [layer.fd_config.model_config.num_key_value_heads] scale_shape = [layer.fd_config.model_config.num_key_value_heads]
if self.cache_quant_config.is_channel_wise: if self.cache_quant_config.is_channel_wise:
scale_shape = [layer.fd_config.model_config.num_key_value_heads, layer.head_dim] scale_shape = [layer.kv_num_heads * layer.head_dim]
layer.cache_k_scale = layer.create_parameter( layer.cache_k_scale = layer.create_parameter(
shape=scale_shape, shape=scale_shape,

3
tests/quantization/test_kv_cache.py
View File

@@ -37,6 +37,7 @@ class MockLayer(nn.Layer):
self.fd_config = get_default_test_fd_config() self.fd_config = get_default_test_fd_config()
self.fd_config.model_config.num_key_value_heads = 1 self.fd_config.model_config.num_key_value_heads = 1
self.head_dim = 1 self.head_dim = 1
self.kv_num_heads = 1
self.prefix = "mock_layer" self.prefix = "mock_layer"
self.cache_k_scale = None self.cache_k_scale = None
self.cache_v_scale = None self.cache_v_scale = None
@@ -77,7 +78,7 @@ class TestKVCacheMethodBase(unittest.TestCase):
method = KVCacheMethodBase(config) method = KVCacheMethodBase(config)
method.create_weights(self.layer) method.create_weights(self.layer)
self.assertEqual(self.layer.cache_k_scale.shape, [1, 1]) self.assertEqual(self.layer.cache_k_scale.shape, [1])
def test_create_weights_int4_zp(self): def test_create_weights_int4_zp(self):
# Test INT4 with zero point # Test INT4 with zero point