maulikanalog commited on
Commit
a721b76
·
verified ·
1 Parent(s): d783039

Upload 2 files

Browse files
Files changed (2) hide show
  1. README (2).md +53 -0
  2. config (3).yaml +66 -0
README (2).md ADDED
@@ -0,0 +1,53 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ tags:
3
+ - text-to-image
4
+ - flux
5
+ - lora
6
+ - diffusers
7
+ - template:sd-lora
8
+ - ai-toolkit
9
+ widget:
10
+ - text: A person in a bustling cafe pareshv
11
+ output:
12
+ url: samples/1738127149163__000003000_0.jpg
13
+ - text: pareshv in tailored Italian suit
14
+ output:
15
+ url: samples/1738127162668__000003000_1.jpg
16
+ - text: pareshv in tailored Italian blue suit in office
17
+ output:
18
+ url: samples/1738127176178__000003000_2.jpg
19
+ base_model: black-forest-labs/FLUX.1-dev
20
+ instance_prompt: pareshv
21
+ license: other
22
+ license_name: flux-1-dev-non-commercial-license
23
+ license_link: https://huggingface.co/black-forest-labs/FLUX.1-dev/blob/main/LICENSE.md
24
+ ---
25
+
26
+ # pareshv
27
+ Model trained with [AI Toolkit by Ostris](https://github.com/ostris/ai-toolkit)
28
+ <Gallery />
29
+
30
+ ## Trigger words
31
+
32
+ You should use `pareshv` to trigger the image generation.
33
+
34
+ ## Download model and use it with ComfyUI, AUTOMATIC1111, SD.Next, Invoke AI, etc.
35
+
36
+ Weights for this model are available in Safetensors format.
37
+
38
+ [Download](/None/tree/main) them in the Files & versions tab.
39
+
40
+ ## Use it with the [🧨 diffusers library](https://github.com/huggingface/diffusers)
41
+
42
+ ```py
43
+ from diffusers import AutoPipelineForText2Image
44
+ import torch
45
+
46
+ pipeline = AutoPipelineForText2Image.from_pretrained('black-forest-labs/FLUX.1-dev', torch_dtype=torch.bfloat16).to('cuda')
47
+ pipeline.load_lora_weights('None', weight_name='pareshv')
48
+ image = pipeline('A person in a bustling cafe pareshv').images[0]
49
+ image.save("my_image.png")
50
+ ```
51
+
52
+ For more details, including weighting, merging and fusing LoRAs, check the [documentation on loading LoRAs in diffusers](https://huggingface.co/docs/diffusers/main/en/using-diffusers/loading_adapters)
53
+
config (3).yaml ADDED
@@ -0,0 +1,66 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ job: extension
2
+ config:
3
+ name: pareshv
4
+ process:
5
+ - type: sd_trainer
6
+ training_folder: output
7
+ device: cuda:0
8
+ network:
9
+ type: lora
10
+ linear: 16
11
+ linear_alpha: 16
12
+ save:
13
+ dtype: float16
14
+ save_every: 10000
15
+ max_step_saves_to_keep: 4
16
+ push_to_hub: true
17
+ hf_private: true
18
+ datasets:
19
+ - folder_path: datasets/67556bbe-bfb5-47e5-a2ad-9533fab0167a
20
+ caption_ext: txt
21
+ caption_dropout_rate: 0.05
22
+ shuffle_tokens: false
23
+ cache_latents_to_disk: true
24
+ resolution:
25
+ - 512
26
+ - 768
27
+ - 1024
28
+ train:
29
+ batch_size: 1
30
+ steps: 3000
31
+ gradient_accumulation_steps: 1
32
+ train_unet: true
33
+ train_text_encoder: false
34
+ gradient_checkpointing: true
35
+ noise_scheduler: flowmatch
36
+ optimizer: adamw8bit
37
+ lr: 0.0004
38
+ ema_config:
39
+ use_ema: true
40
+ ema_decay: 0.99
41
+ dtype: bf16
42
+ skip_first_sample: true
43
+ disable_sampling: false
44
+ model:
45
+ name_or_path: black-forest-labs/FLUX.1-dev
46
+ is_flux: true
47
+ quantize: true
48
+ low_vram: false
49
+ sample:
50
+ sampler: flowmatch
51
+ sample_every: 1500
52
+ width: 1024
53
+ height: 1024
54
+ prompts:
55
+ - A person in a bustling cafe pareshv
56
+ - pareshv in tailored Italian suit
57
+ - pareshv in tailored Italian blue suit in office
58
+ neg: ''
59
+ seed: 42
60
+ walk_seed: true
61
+ guidance_scale: 3.5
62
+ sample_steps: 28
63
+ trigger_word: pareshv
64
+ meta:
65
+ name: pareshv
66
+ version: '1.0'