Feature/batched image generation

[sokovninn]

Added batched image generation

Some benchmarks (NVIDIA L4 24GB GPU):

SDXL-Turbo (~6x speed up):

• 1000 images batch_size=1: 68 min (7.1 GB)
• 1000 images batch_size=2: 37 min (8.1 GB)
• 1000 images batch_size=4: 22 min (11.5 GB)
• 1000 images batch_size=8: 15 min (17.6 GB)
• 1000 images batch_size=16 (max): 11 min (21.2 GB)

The time for 10k images reduced to <2h.

SDXL (almost no speed up for some reason)

• 100 images batch_size=1: 60 min
• 100 images batch_size=4 (max): 55 min

SDXL-Turbo on CPU:

• 100 images batch_size=1: 48 min
• 100 images batch_size=16: 60 min

[github-actions]

☂️ Python Coverage

current status: ✅

Overall Coverage

Lines	Covered	Coverage	Threshold	Status
799	375	47%	0%	🟢

New Files

No new covered files...

Modified Files

File	Coverage	Status
datadreamer/image_generation/clip_image_tester.py	52%	🟢
datadreamer/image_generation/image_generator.py	25%	🟢
datadreamer/image_generation/sdxl_image_generator.py	20%	🟢
datadreamer/image_generation/sdxl_turbo_image_generator.py	26%	🟢
datadreamer/pipelines/generate_dataset_from_scratch.py	43%	🟢
datadreamer/prompt_generation/lm_prompt_generator.py	58%	🟢
datadreamer/prompt_generation/tinyllama_lm_prompt_generator.py	82%	🟢
TOTAL	44%	🟢

updated for commit: 9f12c57 by action🐍

[github-actions]

Test Results

  6 files    6 suites   46m 59s ⏱️
 71 tests  29 ✅  42 💤 0 ❌
426 runs  174 ✅ 252 💤 0 ❌

Results for commit 9f12c57.

[HonzaCuhel]

LGTM