tckm.yan/aha
0
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

update candle version 0.9.2

Browse Source
This commit is contained in:
jhqxxx
2026-02-24 16:34:06 +08:00
parent e19c1d0744
commit 6a73ee04c3
3 changed files with 342 additions and 297 deletions
Download Patch File Download Diff File Expand all files Collapse all files
Cargo.lock
Generated
+325 -293
View File
File diff suppressed because it is too large Load Diff
Cargo.toml
+4 -4
View File
@@ -7,10 +7,10 @@ license = "Apache-2.0"
description = "aha model inference library, now supports Qwen2.5VL, MiniCPM4, VoxCPM, Qwen3VL, DeepSeek-OCR, Hunyuan-OCR, PaddleOCR-VL, VoxCPM1.5, RMBG2.0, GLM-ASR-Nano-2512, Fun-ASR-Nano-2512, Qwen3, Qwen3-ASR"
[dependencies]
candle-core = { version = "0.9.1" }
candle-nn = { version = "0.9.1" }
candle-transformers = { version = "0.9.1" }
candle-flash-attn = { version = "0.9.1", optional = true }
candle-core = { version = "0.9.2" }
candle-nn = { version = "0.9.2" }
candle-transformers = { version = "0.9.2" }
candle-flash-attn = { version = "0.9.2", optional = true }
serde = "1.0.226"
serde_json = "1.0.145"
anyhow = "1.0.100"
src/utils/mod.rs
+13
View File
@@ -210,11 +210,24 @@ pub fn crate_tensor_from_reader<R: std::io::Read>(
reader.read_u32_into::<LittleEndian>(&mut data_t)?;
Ok(Tensor::from_vec(data_t, shape, &Device::Cpu)?)
}
DType::I16 => {
let mut data_t = vec![0i16; elem_count];
reader.read_i16_into::<LittleEndian>(&mut data_t)?;
Ok(Tensor::from_vec(data_t, shape, &Device::Cpu)?)
}
DType::I32 => {
let mut data_t = vec![0i32; elem_count];
reader.read_i32_into::<LittleEndian>(&mut data_t)?;
Ok(Tensor::from_vec(data_t, shape, &Device::Cpu)?)
}
DType::I64 => {
let mut data_t = vec![0i64; elem_count];
reader.read_i64_into::<LittleEndian>(&mut data_t)?;
Ok(Tensor::from_vec(data_t, shape, &Device::Cpu)?)
}
DType::F8E4M3 | DType::F6E2M3 | DType::F6E3M2 | DType::F4 | DType::F8E8M0 => {
Err(anyhow!(format!("UnsupportedDTypeForOp '{:?}'", dtype)))
}
}
}