aditya11997 commited on
Commit
b0371f0
1 Parent(s): 4ea42b7

End of training

Browse files
README.md ADDED
@@ -0,0 +1,50 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+
2
+ ---
3
+ license: creativeml-openrail-m
4
+ base_model: kandinsky-community/kandinsky-2-2-prior
5
+ datasets:
6
+ - kbharat7/DogChestXrayDatasetNew
7
+ tags:
8
+ - kandinsky
9
+ - text-to-image
10
+ - diffusers
11
+ - diffusers-training
12
+ inference: true
13
+ ---
14
+
15
+ # Finetuning - aditya11997/test_prior
16
+
17
+ This pipeline was finetuned from **kandinsky-community/kandinsky-2-2-prior** on the **kbharat7/DogChestXrayDatasetNew** dataset. Below are some example images generated with the finetuned pipeline using the following prompts: ['dogxraysmall']:
18
+
19
+ ![val_imgs_grid](./val_imgs_grid.png)
20
+
21
+
22
+ ## Pipeline usage
23
+
24
+ You can use the pipeline like so:
25
+
26
+ ```python
27
+ from diffusers import DiffusionPipeline
28
+ import torch
29
+
30
+ pipe_prior = DiffusionPipeline.from_pretrained("aditya11997/test_prior", torch_dtype=torch.float16)
31
+ pipe_t2i = DiffusionPipeline.from_pretrained("kandinsky-community/kandinsky-2-2-decoder", torch_dtype=torch.float16)
32
+ prompt = "dogxraysmall"
33
+ image_embeds, negative_image_embeds = pipe_prior(prompt, guidance_scale=1.0).to_tuple()
34
+ image = pipe_t2i(image_embeds=image_embeds, negative_image_embeds=negative_image_embeds).images[0]
35
+ image.save("my_image.png")
36
+ ```
37
+
38
+ ## Training info
39
+
40
+ These are the key hyperparameters used during training:
41
+
42
+ * Epochs: 1
43
+ * Learning rate: 1e-05
44
+ * Batch size: 1
45
+ * Gradient accumulation steps: 4
46
+ * Image resolution: 768
47
+ * Mixed-precision: None
48
+
49
+
50
+ More information on all the CLI arguments and the environment are available on your [`wandb` run page](https://wandb.ai/aditya11997/text2image-fine-tune/runs/9j7m0fr8).
image_encoder/config.json ADDED
@@ -0,0 +1,23 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "kandinsky-community/kandinsky-2-2-prior",
3
+ "architectures": [
4
+ "CLIPVisionModelWithProjection"
5
+ ],
6
+ "attention_dropout": 0.0,
7
+ "dropout": 0.0,
8
+ "hidden_act": "gelu",
9
+ "hidden_size": 1664,
10
+ "image_size": 224,
11
+ "initializer_factor": 1.0,
12
+ "initializer_range": 0.02,
13
+ "intermediate_size": 8192,
14
+ "layer_norm_eps": 1e-05,
15
+ "model_type": "clip_vision_model",
16
+ "num_attention_heads": 16,
17
+ "num_channels": 3,
18
+ "num_hidden_layers": 48,
19
+ "patch_size": 14,
20
+ "projection_dim": 1280,
21
+ "torch_dtype": "float16",
22
+ "transformers_version": "4.38.2"
23
+ }
image_encoder/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:dabf125eaf7f339da6b900a7a14f8190f02913b7d46cab80e09cd0755c4ce39c
3
+ size 3689910512
image_processor/preprocessor_config.json ADDED
@@ -0,0 +1,27 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "crop_size": {
3
+ "height": 224,
4
+ "width": 224
5
+ },
6
+ "do_center_crop": true,
7
+ "do_convert_rgb": true,
8
+ "do_normalize": true,
9
+ "do_rescale": true,
10
+ "do_resize": true,
11
+ "image_mean": [
12
+ 0.48145466,
13
+ 0.4578275,
14
+ 0.40821073
15
+ ],
16
+ "image_processor_type": "CLIPImageProcessor",
17
+ "image_std": [
18
+ 0.26862954,
19
+ 0.26130258,
20
+ 0.27577711
21
+ ],
22
+ "resample": 3,
23
+ "rescale_factor": 0.00392156862745098,
24
+ "size": {
25
+ "shortest_edge": 224
26
+ }
27
+ }
logs/text2image-fine-tune/1709712401.761387/events.out.tfevents.1709712401.777f50584738.1857.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:38b63accc930690c735540783e5103df541e0de660ea49ac001e2afde5b57fc1
3
+ size 2030
logs/text2image-fine-tune/1709712401.7630112/hparams.yml ADDED
@@ -0,0 +1,41 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ adam_beta1: 0.9
2
+ adam_beta2: 0.999
3
+ adam_epsilon: 1.0e-08
4
+ adam_weight_decay: 0.0
5
+ allow_tf32: false
6
+ cache_dir: null
7
+ caption_column: text
8
+ checkpointing_steps: 500
9
+ checkpoints_total_limit: 3
10
+ dataloader_num_workers: 0
11
+ dataset_config_name: null
12
+ dataset_name: kbharat7/DogChestXrayDatasetNew
13
+ gradient_accumulation_steps: 4
14
+ hub_model_id: null
15
+ hub_token: null
16
+ image_column: image
17
+ learning_rate: 1.0e-05
18
+ local_rank: -1
19
+ logging_dir: logs
20
+ lr_scheduler: constant
21
+ lr_warmup_steps: 500
22
+ max_grad_norm: 1.0
23
+ max_train_samples: null
24
+ max_train_steps: 10
25
+ mixed_precision: null
26
+ num_train_epochs: 1
27
+ output_dir: test_prior
28
+ pretrained_decoder_model_name_or_path: kandinsky-community/kandinsky-2-2-decoder
29
+ pretrained_prior_model_name_or_path: kandinsky-community/kandinsky-2-2-prior
30
+ push_to_hub: true
31
+ report_to: tensorboard
32
+ resolution: 768
33
+ resume_from_checkpoint: null
34
+ seed: null
35
+ snr_gamma: null
36
+ tracker_project_name: text2image-fine-tune
37
+ train_batch_size: 1
38
+ train_data_dir: null
39
+ use_8bit_adam: false
40
+ use_ema: false
41
+ validation_epochs: 5
logs/text2image-fine-tune/events.out.tfevents.1709712401.777f50584738.1857.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:042af7d775f34c27d5f6d7317b9cdf0344267555aa973722f59a1c78f6bbb81f
3
+ size 336100
model_index.json ADDED
@@ -0,0 +1,28 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_class_name": "KandinskyV22PriorPipeline",
3
+ "_diffusers_version": "0.27.0.dev0",
4
+ "image_encoder": [
5
+ "transformers",
6
+ "CLIPVisionModelWithProjection"
7
+ ],
8
+ "image_processor": [
9
+ "transformers",
10
+ "CLIPImageProcessor"
11
+ ],
12
+ "prior": [
13
+ "diffusers",
14
+ "PriorTransformer"
15
+ ],
16
+ "scheduler": [
17
+ "diffusers",
18
+ "UnCLIPScheduler"
19
+ ],
20
+ "text_encoder": [
21
+ "transformers",
22
+ "CLIPTextModelWithProjection"
23
+ ],
24
+ "tokenizer": [
25
+ "transformers",
26
+ "CLIPTokenizer"
27
+ ]
28
+ }
prior/config.json ADDED
@@ -0,0 +1,20 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_class_name": "PriorTransformer",
3
+ "_diffusers_version": "0.27.0.dev0",
4
+ "_name_or_path": "kandinsky-community/kandinsky-2-2-prior",
5
+ "added_emb_type": "prd",
6
+ "additional_embeddings": 4,
7
+ "attention_head_dim": 64,
8
+ "clip_embed_dim": null,
9
+ "dropout": 0.0,
10
+ "embedding_dim": 1280,
11
+ "embedding_proj_dim": null,
12
+ "embedding_proj_norm_type": null,
13
+ "encoder_hid_proj_type": "linear",
14
+ "norm_in_type": null,
15
+ "num_attention_heads": 32,
16
+ "num_embeddings": 77,
17
+ "num_layers": 20,
18
+ "time_embed_act_fn": "silu",
19
+ "time_embed_dim": null
20
+ }
prior/diffusion_pytorch_model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:28007025d5506a871d345ef898846d449ad6528cfd95e98f00ed87efa06d63b5
3
+ size 4104940968
scheduler/scheduler_config.json ADDED
@@ -0,0 +1,10 @@
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_class_name": "UnCLIPScheduler",
3
+ "_diffusers_version": "0.27.0.dev0",
4
+ "beta_schedule": "squaredcos_cap_v2",
5
+ "clip_sample": true,
6
+ "clip_sample_range": 10.0,
7
+ "num_train_timesteps": 1000,
8
+ "prediction_type": "sample",
9
+ "variance_type": "fixed_small_log"
10
+ }
text_encoder/config.json ADDED
@@ -0,0 +1,25 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "kandinsky-community/kandinsky-2-2-prior",
3
+ "architectures": [
4
+ "CLIPTextModelWithProjection"
5
+ ],
6
+ "attention_dropout": 0.0,
7
+ "bos_token_id": 0,
8
+ "dropout": 0.0,
9
+ "eos_token_id": 2,
10
+ "hidden_act": "gelu",
11
+ "hidden_size": 1280,
12
+ "initializer_factor": 1.0,
13
+ "initializer_range": 0.02,
14
+ "intermediate_size": 5120,
15
+ "layer_norm_eps": 1e-05,
16
+ "max_position_embeddings": 77,
17
+ "model_type": "clip_text_model",
18
+ "num_attention_heads": 20,
19
+ "num_hidden_layers": 32,
20
+ "pad_token_id": 1,
21
+ "projection_dim": 1280,
22
+ "torch_dtype": "float16",
23
+ "transformers_version": "4.38.2",
24
+ "vocab_size": 49408
25
+ }
text_encoder/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ec310df2af79c318e24d20511b601a591ca8cd4f1fce1d8dff822a356bcdb1f4
3
+ size 1389382176
tokenizer/merges.txt ADDED
The diff for this file is too large to render. See raw diff
 
tokenizer/special_tokens_map.json ADDED
@@ -0,0 +1,24 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": {
3
+ "content": "<|startoftext|>",
4
+ "lstrip": false,
5
+ "normalized": true,
6
+ "rstrip": false,
7
+ "single_word": false
8
+ },
9
+ "eos_token": {
10
+ "content": "<|endoftext|>",
11
+ "lstrip": false,
12
+ "normalized": true,
13
+ "rstrip": false,
14
+ "single_word": false
15
+ },
16
+ "pad_token": "<|endoftext|>",
17
+ "unk_token": {
18
+ "content": "<|endoftext|>",
19
+ "lstrip": false,
20
+ "normalized": true,
21
+ "rstrip": false,
22
+ "single_word": false
23
+ }
24
+ }
tokenizer/tokenizer_config.json ADDED
@@ -0,0 +1,30 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_prefix_space": false,
3
+ "added_tokens_decoder": {
4
+ "49406": {
5
+ "content": "<|startoftext|>",
6
+ "lstrip": false,
7
+ "normalized": true,
8
+ "rstrip": false,
9
+ "single_word": false,
10
+ "special": true
11
+ },
12
+ "49407": {
13
+ "content": "<|endoftext|>",
14
+ "lstrip": false,
15
+ "normalized": true,
16
+ "rstrip": false,
17
+ "single_word": false,
18
+ "special": true
19
+ }
20
+ },
21
+ "bos_token": "<|startoftext|>",
22
+ "clean_up_tokenization_spaces": true,
23
+ "do_lower_case": true,
24
+ "eos_token": "<|endoftext|>",
25
+ "errors": "replace",
26
+ "model_max_length": 77,
27
+ "pad_token": "<|endoftext|>",
28
+ "tokenizer_class": "CLIPTokenizer",
29
+ "unk_token": "<|endoftext|>"
30
+ }
tokenizer/vocab.json ADDED
The diff for this file is too large to render. See raw diff
 
val_imgs_grid.png ADDED