Skip to content

Chroma support (pruned Flux model) #696

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 21 commits into from
Jun 29, 2025
Merged

Chroma support (pruned Flux model) #696

Changes from 1 commit
Commits
Show all changes
21 commits
Select commit Hold shift + click to select a range
f7ad456
Chroma: Initial commit (broken output)
stduhpf May 28, 2025
ad39011
Fix small mistake (still broken)
stduhpf May 29, 2025
93ed721
is_chroma
stduhpf May 29, 2025
d20f77f
Reshape before approx
stduhpf May 30, 2025
bb1fe2c
format
stduhpf Apr 10, 2025
f506a63
Fix use_after_free (hopefully)
stduhpf May 30, 2025
836fd72
Chroma: Attention masking (no pad)
stduhpf Jun 1, 2025
67cc996
implement chroma mask padding
stduhpf Jun 2, 2025
55a2686
Use env variable to control chroma padding settings
stduhpf Jun 2, 2025
5a96b09
std::isinf
stduhpf Jun 2, 2025
2bdc847
Use ggml_pad instead of concat with empty tensor
stduhpf Jun 2, 2025
322fad2
Optimise masked attention
stduhpf Jun 2, 2025
42e217d
Chroma: disable guidance by default
stduhpf Jun 2, 2025
4fdedd5
Chroma: Fix t5 chunk length
stduhpf Jun 2, 2025
3238fe3
fix mask with flash attn
Green-Sky Jun 3, 2025
efaa137
Remove deprecated Todos
stduhpf Jun 3, 2025
e22c57c
Only include pad into mask once
stduhpf Jun 4, 2025
00141e5
Chroma/Pixart T5: Disable attention masking by default
stduhpf Jun 14, 2025
60c2468
Merge branch 'master' into chroma
leejet Jun 29, 2025
478992e
use parameters instead of environment variables
leejet Jun 29, 2025
7e400c9
try to unify the code style
leejet Jun 29, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
Prev Previous commit
Next Next commit
Chroma: Fix t5 chunk length
  • Loading branch information
@stduhpf
stduhpf committed Jun 2, 2025
commit 4fdedd5ac59bf0d5e4d197077eed53929fbe35e6
31 changes: 16 additions & 15 deletions conditioner.hpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1004,6 +1004,7 @@ struct FluxCLIPEmbedder : public Conditioner {
T5UniGramTokenizer t5_tokenizer;
std::shared_ptr<CLIPTextModelRunner> clip_l;
std::shared_ptr<T5Runner> t5;
size_t chunk_len = 256;

FluxCLIPEmbedder(ggml_backend_t backend,
std::map<std::string, enum ggml_type>& tensor_types,
Expand Down Expand Up @@ -1109,7 +1110,6 @@ struct FluxCLIPEmbedder : public Conditioner {
struct ggml_tensor* pooled = NULL; // [768,]
std::vector<float> hidden_states_vec;

size_t chunk_len = 256;
size_t chunk_count = t5_tokens.size() / chunk_len;
for (int chunk_idx = 0; chunk_idx < chunk_count; chunk_idx++) {
// clip_l
Expand Down Expand Up @@ -1196,7 +1196,7 @@ struct FluxCLIPEmbedder : public Conditioner {
int height,
int adm_in_channels = -1,
bool force_zero_embeddings = false) {
auto tokens_and_weights = tokenize(text, 256, true);
auto tokens_and_weights = tokenize(text, chunk_len, true);
return get_learned_condition_common(work_ctx, n_threads, tokens_and_weights, clip_skip, force_zero_embeddings);
}

Expand All @@ -1221,6 +1221,7 @@ struct FluxCLIPEmbedder : public Conditioner {
struct PixArtCLIPEmbedder : public Conditioner {
T5UniGramTokenizer t5_tokenizer;
std::shared_ptr<T5Runner> t5;
size_t chunk_len = 512;

PixArtCLIPEmbedder(ggml_backend_t backend,
std::map<std::string, enum ggml_type>& tensor_types,
Expand Down Expand Up @@ -1304,8 +1305,18 @@ struct PixArtCLIPEmbedder : public Conditioner {

std::vector<float> hidden_states_vec;

size_t chunk_len = 256;
size_t chunk_count = t5_tokens.size() / chunk_len;

bool use_mask = true;
const char* SD_CHROMA_USE_T5_MASK = getenv("SD_CHROMA_USE_T5_MASK");
if (SD_CHROMA_USE_T5_MASK != nullptr) {
std::string sd_chroma_use_t5_mask_str = SD_CHROMA_USE_T5_MASK;
if (sd_chroma_use_t5_mask_str == "OFF" || sd_chroma_use_t5_mask_str == "FALSE") {
use_mask = false;
} else if (sd_chroma_use_t5_mask_str != "ON" && sd_chroma_use_t5_mask_str != "TRUE") {
LOG_WARN("SD_CHROMA_USE_T5_MASK environment variable has unexpected value. Assuming default (\"ON\"). (Expected \"ON\"/\"TRUE\" or\"OFF\"/\"FALSE\", got \"%s\")", SD_CHROMA_USE_T5_MASK);
}
}
for (int chunk_idx = 0; chunk_idx < chunk_count; chunk_idx++) {
// t5
std::vector<int> chunk_tokens(t5_tokens.begin() + chunk_idx * chunk_len,
Expand All @@ -1316,17 +1327,7 @@ struct PixArtCLIPEmbedder : public Conditioner {
t5_attn_mask_vec.begin() + (chunk_idx + 1) * chunk_len);

auto input_ids = vector_to_ggml_tensor_i32(work_ctx, chunk_tokens);
auto t5_attn_mask_chunk = vector_to_ggml_tensor(work_ctx, chunk_mask);

const char* SD_CHROMA_USE_T5_MASK = getenv("SD_CHROMA_USE_T5_MASK");
if (SD_CHROMA_USE_T5_MASK != nullptr) {
std::string sd_chroma_use_t5_mask_str = SD_CHROMA_USE_T5_MASK;
if (sd_chroma_use_t5_mask_str == "OFF" || sd_chroma_use_t5_mask_str == "FALSE") {
t5_attn_mask_chunk = NULL;
} else if (sd_chroma_use_t5_mask_str != "ON" && sd_chroma_use_t5_mask_str != "TRUE") {
LOG_WARN("SD_CHROMA_USE_T5_MASK environment variable has unexpected value. Assuming default (\"ON\"). (Expected \"ON\"/\"TRUE\" or\"OFF\"/\"FALSE\", got \"%s\")", SD_CHROMA_USE_T5_MASK);
}
}
auto t5_attn_mask_chunk = use_mask ? vector_to_ggml_tensor(work_ctx, chunk_mask) : NULL;

t5->compute(n_threads,
input_ids,
Expand Down Expand Up @@ -1384,7 +1385,7 @@ struct PixArtCLIPEmbedder : public Conditioner {
int height,
int adm_in_channels = -1,
bool force_zero_embeddings = false) {
auto tokens_and_weights = tokenize(text, 512, true);
auto tokens_and_weights = tokenize(text, chunk_len, true);
return get_learned_condition_common(work_ctx, n_threads, tokens_and_weights, clip_skip, force_zero_embeddings);
}

Expand Down
Loading