DarthAffe/stable-diffusion.cpp
1
0 Fork 0
mirror of https://github.com/leejet/stable-diffusion.cpp.git synced 2026-11-12 11:56:36 +00:00
Code Issues Packages Projects Releases Wiki Activity

feat: add Iris-3B text-to-image support (#2117)

Browse Source
This commit is contained in:
leejet authored and GitHub committed 2026-10-10 01:39:49 +08:00
1 parent f2f168fa70
commit f89d9b13d7
10 files changed
+565 -16
Download Patch File Download Diff File Expand all files Collapse all files

No files matched your search

src/pipeline/diffusion_engine.cpp
+7 -5
View File
@@ -108,6 +108,7 @@ const char* model_version_to_str[] = {
"PixArt",
"Ming-Image",
"Z-Image L2P",
"Iris-3B",
};
static_assert(VERSION_COUNT == sizeof(model_version_to_str) / sizeof(model_version_to_str[0]),
@@ -1382,7 +1383,8 @@ bool StableDiffusionGGML::build_denoiser() {
sd_version_is_llada_image(version) ||
sd_version_is_boogu_image(version) ||
sd_version_is_pid(version) ||
sd_version_is_ideogram4(version)) {
sd_version_is_ideogram4(version) ||
version == VERSION_IRIS) {
pred_type = FLOW_PRED;
if (sd_version_is_wan(version)) {
default_flow_shift = 5.f;
@@ -1390,7 +1392,7 @@ bool StableDiffusionGGML::build_denoiser() {
default_flow_shift = 7.f;
} else if (sd_version_is_minimax_h3(version)) {
default_flow_shift = 12.f;
} else if (sd_version_is_ernie_image(version)) {
} else if (sd_version_is_ernie_image(version) || version == VERSION_IRIS) {
default_flow_shift = 4.f;
} else if (sd_version_is_pid(version)) {
default_flow_shift = 1.5f;
@@ -2765,7 +2767,7 @@ int StableDiffusionGGML::get_diffusion_model_down_factor() {
if (sd_version_is_dit(version)) {
if (sd_version_is_sensenova_u1(version)) {
down_factor = 32;
} else if (sd_version_is_z_image_l2p(version)) {
} else if (sd_version_is_z_image_l2p(version) || version == VERSION_IRIS) {
down_factor = 16;
} else if (version == VERSION_QWEN_IMAGE_2_1 || version == VERSION_MING_IMAGE || sd_version_is_wan(version) || sd_version_is_lingbot_video(version) || sd_version_is_minimax_h3(version) || sd_version_is_pixart(version)) {
down_factor = 2;
@@ -2799,7 +2801,7 @@ int StableDiffusionGGML::get_latent_channel() {
latent_channel = 3;
} else if (sd_version_is_z_image_l2p(version)) {
latent_channel = 3;
} else if (sd_version_is_pid(version)) {
} else if (sd_version_is_pid(version) || version == VERSION_IRIS) {
latent_channel = 3;
} else if (sd_version_is_sefi_image(version)) {
latent_channel = 144;
@@ -2900,7 +2902,7 @@ sd::Tensor<float> StableDiffusionGGML::encode_first_stage(const sd::Tensor<float
}
sd::Tensor<float> StableDiffusionGGML::decode_first_stage(const sd::Tensor<float>& x, bool decode_video) {
if (sd_version_is_pid(version) || sd_version_is_minit2i(version)) {
if (sd_version_is_pid(version) || sd_version_is_minit2i(version) || version == VERSION_IRIS) {
return sd::ops::clamp((x + 1.f) * 0.5f, 0.0f, 1.0f);
}
auto latents = first_stage_model->diffusion_to_vae_latents(x);
src/pipeline/model_builders.cpp
+12 -1
View File
@@ -18,6 +18,7 @@
#include "model/diffusion/hidream_o1.hpp"
#include "model/diffusion/hunyuan.hpp"
#include "model/diffusion/ideogram4.hpp"
#include "model/diffusion/iris.hpp"
#include "model/diffusion/krea2.hpp"
#include "model/diffusion/lens.hpp"
#include "model/diffusion/lingbot_video.hpp"
@@ -456,6 +457,11 @@ namespace sd::model_builders {
tensor_storage_map,
"model.diffusion_model",
weight_manager);
} else if (version == VERSION_IRIS) {
result.conditioner = std::make_shared<LLMEmbedder>(ctx.backends.runtime_backend(SDBackendModule::TE),
tensor_storage_map, version, "", false, weight_manager, tokenizers);
result.diffusion = std::make_shared<Iris::IrisRunner>(ctx.backends.runtime_backend(SDBackendModule::DIFFUSION),
tensor_storage_map, "model.diffusion_model", weight_manager);
} else { // SD1.x SD2.x SDXL
std::map<std::string, std::string> embbeding_map;
for (uint32_t i = 0; i < sd_ctx_params->embedding_count; i++) {
@@ -626,7 +632,12 @@ namespace sd::model_builders {
}
};
if (version == VERSION_CHROMA_RADIANCE || version == VERSION_HIDREAM_O1 || sd_version_is_minit2i(version) || sd_version_is_sensenova_u1(version) || sd_version_is_z_image_l2p(version)) {
if (version == VERSION_CHROMA_RADIANCE ||
version == VERSION_HIDREAM_O1 ||
sd_version_is_minit2i(version) ||
sd_version_is_sensenova_u1(version) ||
sd_version_is_z_image_l2p(version) ||
version == VERSION_IRIS) {
LOG_INFO("using FakeVAE");
result.vae = std::make_shared<FakeVAE>(version,
ctx.backends.runtime_backend(SDBackendModule::VAE),