oxigaf-diffusion 0.1.0

Multi-view diffusion model inference for GAF
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51

[package]
name = "oxigaf-diffusion"
version.workspace = true
edition.workspace = true
authors.workspace = true
license.workspace = true
repository.workspace = true
description = "Multi-view diffusion model inference for GAF"
readme = "README.md"
keywords = ["diffusion", "multi-view", "ai", "deep-learning", "generative"]
categories = ["science", "multimedia", "algorithms"]

[dependencies]
candle-core = { workspace = true }
candle-nn = { workspace = true }
candle-transformers = { workspace = true }
safetensors = { workspace = true }
image = { workspace = true }
thiserror = { workspace = true }
tracing = { workspace = true }

oxigaf-flame = { version = "0.1.0", path = "../oxigaf-flame" }

[dev-dependencies]
proptest = { workspace = true }
approx = { workspace = true }
candle-core = { workspace = true }
candle-nn = { workspace = true }
criterion = { workspace = true }

[[bench]]
name = "diffusion_bench"
harness = false

[[bench]]
name = "flash_attention_bench"
harness = false

[features]
# Default to CPU-only for compatibility on all platforms
default = ["flash_attention"]
# Flash attention for memory-efficient O(N) attention
flash_attention = []
# Mixed-precision inference (FP16/BF16 for reduced memory usage)
# Note: Implementation is planned - currently a placeholder
mixed_precision = []
# NOTE: Platform-specific GPU/BLAS backends (accelerate, metal, cuda) have been
# removed from this crate. To enable them, configure candle-core directly:
#   candle-core = { version = "0.9", features = ["accelerate"] }  # macOS BLAS
#   candle-core = { version = "0.9", features = ["metal"] }       # macOS GPU
#   candle-core = { version = "0.9", features = ["cuda"] }        # NVIDIA GPU