Skip to content

Ig/diffuser 0.34.0 #2152

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
regisss merged 2 commits into from
Jul 17, 2025
Merged

Ig/diffuser 0.34.0 #2152

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
cb385a6
initial commit
Jul 15, 2025
8102ccf
fea(): fixed the sd training examples for lazy mode
Jul 16, 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
8 changes: 4 additions & 4 deletions examples/stable-diffusion/training/README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -434,7 +434,7 @@ We can use the same `dog` dataset for the following example.

To launch SD3 LoRA training on a single Gaudi card, use:
```bash
python train_dreambooth_lora_sd3.py \
PT_HPU_LAZY_MODE=1 python train_dreambooth_lora_sd3.py \
--pretrained_model_name_or_path="stabilityai/stable-diffusion-3-medium-diffusers" \
--dataset_name="dog" \
--instance_prompt="a photo of sks dog" \
Expand Down Expand Up @@ -469,7 +469,7 @@ with `python ../../gaudi_spawn.py --world_size <num-HPUs> train_text_to_image_sd

After training completes, you could directly use `text_to_image_generation.py` sample for inference as follows:
```bash
python ../text_to_image_generation.py \
PT_HPU_LAZY_MODE=1 python ../text_to_image_generation.py \
--model_name_or_path "stabilityai/stable-diffusion-3-medium-diffusers" \
--lora_id dog_lora_sd3 \
--prompts "A picture of a sks dog in a bucket" \
Expand All @@ -491,7 +491,7 @@ We can use the `dog` dataset for the following example.

To launch SD3 full model training on single Gaudi card, use:
```bash
python train_text_to_image_sd3.py \
PT_HPU_LAZY_MODE=1 python train_text_to_image_sd3.py \
--pretrained_model_name_or_path="stabilityai/stable-diffusion-3-medium-diffusers" \
--dataset_name="dog" \
--instance_prompt="a photo of sks dog" \
Expand Down Expand Up @@ -525,7 +525,7 @@ with `python ../../gaudi_spawn.py --world_size <num-HPUs> train_text_to_image_sd

After training completes, you could directly use `text_to_image_generation.py` sample for inference as follows:
```bash
python ../text_to_image_generation.py \
PT_HPU_LAZY_MODE=1 python ../text_to_image_generation.py \
--model_name_or_path "dog_ft_sd3" \
--prompts "A picture of a sks dog in a bucket" \
--scheduler flow_match_euler_discrete \
Expand Down
2 changes: 1 addition & 1 deletion setup.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -33,7 +33,7 @@
"optimum ~= 1.25",
"torch",
"accelerate >= 1.7.0",
"diffusers >= 0.33.1, < 0.33.2",
"diffusers >= 0.34.0, < 0.34.1",
"huggingface_hub[hf_xet] >= 0.24.7",
"sentence-transformers == 3.3.1",
]
Expand Down
Loading