Replace adapter-only release with complete standalone model

Adds fused full Diffusers pipeline, standalone safetensors checkpoint, beginner launcher, bilingual documentation, prompts, examples, and independent tests.

.gitattributes

@@ -56,3 +56,4 @@ examples/prompt_control/05_woman_base.png filter=lfs diff=lfs merge=lfs -text
 examples/prompt_control/06_dog_base.png filter=lfs diff=lfs merge=lfs -text
 examples/prompt_control/07_woman_trigger_03.png filter=lfs diff=lfs merge=lfs -text
 examples/prompt_control/08_woman_trigger_04.png filter=lfs diff=lfs merge=lfs -text
+examples/full_model_smoke_test.png filter=lfs diff=lfs merge=lfs -text

InkNoir_Core_Full.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6b26172bac01af58119c5d229d6a03197e1e5a9de497d67613dafeea1ad8ff2d
+size 2132625150

MODEL_LICENSE_NOTICE.md

@@ -0,0 +1,15 @@
+# InkNoir Core license notice
+
+## English
+
+The full checkpoint is a fused derivative of [Lykon/DreamShaper 8](https://huggingface.co/Lykon/dreamshaper-8), whose model card identifies its license as **CreativeML Open RAIL-M** (`creativeml-openrail-m`). Users must comply with that license, including its use-based restrictions, attribution requirements and downstream-notice requirements. The original model card is the authoritative source for the current terms.
+
+This repository does not grant permission to ignore the base model’s restrictions. The InkNoir additions are provided as-is. Training-image provenance and any rights needed for redistribution of derived weights remain the responsibility of the publisher.
+
+License reference: https://github.com/CompVis/stable-diffusion/blob/main/LICENSE
+
+## Türkçe
+
+Bu tam checkpoint, model kartında **CreativeML Open RAIL-M** (`creativeml-openrail-m`) lisansıyla belirtilen [Lykon/DreamShaper 8](https://huggingface.co/Lykon/dreamshaper-8) modelinden türetilmiştir. Kullanıcılar bu lisansın kullanım kısıtlarına, atıf şartlarına ve türetilmiş modeller için bildirim şartlarına uymalıdır. Güncel ve bağlayıcı şartlar için orijinal model kartı ve lisans metni esas alınmalıdır.
+
+Bu repo taban modelin kısıtlarını kaldırmaz. InkNoir eklemeleri olduğu gibi sunulur. Eğitim görsellerinin kaynağı ve türetilmiş ağırlıkların dağıtımı için gerekli hakların doğrulanması yayıncının sorumluluğundadır.

PUBLISH_CHECKLIST.md

@@ -1,18 +1,13 @@
-# InkNoir Core release checklist
+# InkNoir Core full release checklist
 
-- [x] Adapter file is valid `safetensors` and contains model metadata.
-- [x] Base model is recorded as `Lykon/dreamshaper-8`.
-- [x] Trigger word and recommended inference settings are documented.
-- [x] Diffusers and Automatic1111/ComfyUI usage examples are documented.
-- [x] English and Turkish setup/use instructions are documented.
-- [x] Six generated example images are included.
-- [x] Each example has its English prompt, Turkish translation, seed and settings in `examples/prompts.md`.
-- [x] System requirements, memory guidance, img2img and troubleshooting sections are documented.
-- [x] Limitations are stated: this is a LoRA, not a standalone model or video model.
-- [x] A sample generated directly with the packaged adapter is included.
-- [x] Packaged adapter SHA-256: `f1c0d425b4a8950b2f212044a7e3b538b17618a5dccec026adb6e424b0463ef0`.
-- [ ] Confirm the Hugging Face account/repository owner name before pushing.
-- [ ] Confirm the redistribution rights and attribution for every locally curated training image.
-- [ ] Review the DreamShaper 8 license terms at publish time.
-
-The last three checks are intentionally left for the owner before public release. The package must not imply that local training-image provenance has been independently cleared when it has not.
+- [x] Complete Diffusers pipeline included at repository root.
+- [x] Standalone half-precision `InkNoir_Core_Full.safetensors` included.
+- [x] Optional unfused adapter included under `adapters/`.
+- [x] Full pipeline loaded locally without a LoRA and generated a smoke-test image.
+- [x] No `lora_*` keys remain in the fused UNet or single-file checkpoint.
+- [x] Beginner launcher `run_inknoir.py` included.
+- [x] `requirements.txt` included.
+- [x] English/Turkish setup and usage instructions included.
+- [x] Main gallery, prompt catalog, independent benchmark and prompt-control tests included.
+- [x] DreamShaper 8 is attributed as the base model and CreativeML Open RAIL-M is identified.
+- [ ] Publisher must verify training-image redistribution rights before claiming a fully cleared dataset provenance.

model_index.json

@@ -0,0 +1,38 @@
+{
+  "_class_name": "StableDiffusionPipeline",
+  "_diffusers_version": "0.41.0.dev0",
+  "_name_or_path": "Lykon/dreamshaper-8",
+  "feature_extractor": [
+    null,
+    null
+  ],
+  "image_encoder": [
+    null,
+    null
+  ],
+  "requires_safety_checker": false,
+  "safety_checker": [
+    null,
+    null
+  ],
+  "scheduler": [
+    "diffusers",
+    "DEISMultistepScheduler"
+  ],
+  "text_encoder": [
+    "transformers",
+    "CLIPTextModel"
+  ],
+  "tokenizer": [
+    "transformers",
+    "CLIPTokenizer"
+  ],
+  "unet": [
+    "diffusers",
+    "UNet2DConditionModel"
+  ],
+  "vae": [
+    "diffusers",
+    "AutoencoderKL"
+  ]
+}

requirements.txt

@@ -0,0 +1,8 @@
+torch
+torchvision
+diffusers
+transformers
+accelerate
+safetensors
+pillow
+sentencepiece

run_inknoir.py

@@ -0,0 +1,63 @@
+#!/usr/bin/env python3
+"""Beginner-friendly InkNoir Core full-model launcher."""
+
+import argparse
+from pathlib import Path
+
+import torch
+from diffusers import StableDiffusionPipeline
+
+
+def main():
+    root = Path(__file__).resolve().parent
+    parser = argparse.ArgumentParser(description="Generate one image with the full InkNoir Core model.")
+    parser.add_argument("--prompt", default=None, help="English or translated prompt")
+    parser.add_argument("--negative", default="low quality, blurry, deformed, bad anatomy, text, watermark")
+    parser.add_argument("--output", default="inknoir_output.png")
+    parser.add_argument("--steps", type=int, default=28)
+    parser.add_argument("--cfg", type=float, default=7.0)
+    parser.add_argument("--seed", type=int, default=20260920)
+    parser.add_argument("--width", type=int, default=512)
+    parser.add_argument("--height", type=int, default=512)
+    parser.add_argument("--model", default=str(root), help="Local model directory or Hugging Face repo ID")
+    args = parser.parse_args()
+
+    prompt = args.prompt or input("Prompt / İstek: ").strip()
+    if not prompt:
+        raise SystemExit("A prompt is required / Prompt gerekli.")
+
+    if torch.cuda.is_available():
+        device, dtype = "cuda", torch.float16
+    elif getattr(torch.backends, "mps", None) and torch.backends.mps.is_available():
+        device, dtype = "mps", torch.float32
+    else:
+        device, dtype = "cpu", torch.float32
+
+    print(f"InkNoir Core yükleniyor: {args.model}")
+    print(f"Cihaz: {device} | Boyut: {args.width}x{args.height} | Adım: {args.steps}")
+    pipe = StableDiffusionPipeline.from_pretrained(
+        args.model,
+        torch_dtype=dtype,
+        safety_checker=None,
+        requires_safety_checker=False,
+    ).to(device)
+    pipe.enable_attention_slicing()
+    if hasattr(pipe, "enable_vae_slicing"):
+        pipe.enable_vae_slicing()
+
+    generator = torch.Generator(device=device).manual_seed(args.seed)
+    image = pipe(
+        prompt=prompt,
+        negative_prompt=args.negative,
+        width=args.width,
+        height=args.height,
+        num_inference_steps=args.steps,
+        guidance_scale=args.cfg,
+        generator=generator,
+    ).images[0]
+    image.save(args.output)
+    print(f"Kaydedildi / Saved: {args.output}")
+
+
+if __name__ == "__main__":
+    main()

scheduler/scheduler_config.json

@@ -0,0 +1,29 @@
+{
+  "_class_name": "DEISMultistepScheduler",
+  "_diffusers_version": "0.41.0.dev0",
+  "algorithm_type": "deis",
+  "beta_end": 0.012,
+  "beta_schedule": "scaled_linear",
+  "beta_start": 0.00085,
+  "clip_sample": false,
+  "dynamic_thresholding_ratio": 0.995,
+  "flow_shift": 1.0,
+  "lower_order_final": true,
+  "num_train_timesteps": 1000,
+  "prediction_type": "epsilon",
+  "sample_max_value": 1.0,
+  "set_alpha_to_one": false,
+  "skip_prk_steps": true,
+  "solver_order": 2,
+  "solver_type": "logrho",
+  "steps_offset": 1,
+  "thresholding": false,
+  "time_shift_type": "exponential",
+  "timestep_spacing": "leading",
+  "trained_betas": null,
+  "use_beta_sigmas": false,
+  "use_dynamic_shifting": false,
+  "use_exponential_sigmas": false,
+  "use_flow_sigmas": false,
+  "use_karras_sigmas": false
+}

text_encoder/config.json

@@ -0,0 +1,24 @@
+{
+  "architectures": [
+    "CLIPTextModel"
+  ],
+  "attention_dropout": 0.0,
+  "bos_token_id": 0,
+  "dropout": 0.0,
+  "dtype": "float16",
+  "eos_token_id": 2,
+  "hidden_act": "quick_gelu",
+  "hidden_size": 768,
+  "initializer_factor": 1.0,
+  "initializer_range": 0.02,
+  "intermediate_size": 3072,
+  "layer_norm_eps": 1e-05,
+  "max_position_embeddings": 77,
+  "model_type": "clip_text_model",
+  "num_attention_heads": 12,
+  "num_hidden_layers": 12,
+  "pad_token_id": 1,
+  "projection_dim": 768,
+  "transformers_version": "5.17.0",
+  "vocab_size": 49408
+}

text_encoder/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fda1f312e49f667179f53878e6ac717d8f37d6119cc5adb76fba88895b94d0a7
+size 246144152

tokenizer/tokenizer_config.json

@@ -0,0 +1,15 @@
+{
+  "add_prefix_space": false,
+  "backend": "tokenizers",
+  "bos_token": "<|startoftext|>",
+  "clean_up_tokenization_spaces": true,
+  "do_lower_case": true,
+  "eos_token": "<|endoftext|>",
+  "errors": "replace",
+  "is_local": true,
+  "local_files_only": true,
+  "model_max_length": 77,
+  "pad_token": "<|endoftext|>",
+  "tokenizer_class": "CLIPTokenizer",
+  "unk_token": "<|endoftext|>"
+}

unet/config.json

@@ -0,0 +1,68 @@
+{
+  "_class_name": "UNet2DConditionModel",
+  "_diffusers_version": "0.41.0.dev0",
+  "_name_or_path": "/home/yunustas/InkNoir-Core/hf_cache/hub/models--Lykon--dreamshaper-8/snapshots/a7e52b98680b1ba8ff7bce97c7f9f2e2e5337917/unet",
+  "act_fn": "silu",
+  "addition_embed_type": null,
+  "addition_embed_type_num_heads": 64,
+  "addition_time_embed_dim": null,
+  "attention_head_dim": 8,
+  "attention_type": "default",
+  "block_out_channels": [
+    320,
+    640,
+    1280,
+    1280
+  ],
+  "center_input_sample": false,
+  "class_embed_type": null,
+  "class_embeddings_concat": false,
+  "conv_in_kernel": 3,
+  "conv_out_kernel": 3,
+  "cross_attention_dim": 768,
+  "cross_attention_norm": null,
+  "down_block_types": [
+    "CrossAttnDownBlock2D",
+    "CrossAttnDownBlock2D",
+    "CrossAttnDownBlock2D",
+    "DownBlock2D"
+  ],
+  "downsample_padding": 1,
+  "dropout": 0.0,
+  "dual_cross_attention": false,
+  "encoder_hid_dim": null,
+  "encoder_hid_dim_type": null,
+  "flip_sin_to_cos": true,
+  "freq_shift": 0,
+  "in_channels": 4,
+  "layers_per_block": 2,
+  "mid_block_only_cross_attention": null,
+  "mid_block_scale_factor": 1,
+  "mid_block_type": "UNetMidBlock2DCrossAttn",
+  "norm_eps": 1e-05,
+  "norm_num_groups": 32,
+  "num_attention_heads": null,
+  "num_class_embeds": null,
+  "only_cross_attention": false,
+  "out_channels": 4,
+  "projection_class_embeddings_input_dim": null,
+  "resnet_out_scale_factor": 1.0,
+  "resnet_skip_time_act": false,
+  "resnet_time_scale_shift": "default",
+  "reverse_transformer_layers_per_block": null,
+  "sample_size": 64,
+  "time_cond_proj_dim": null,
+  "time_embedding_act_fn": null,
+  "time_embedding_dim": null,
+  "time_embedding_type": "positional",
+  "timestep_post_act": null,
+  "transformer_layers_per_block": 1,
+  "up_block_types": [
+    "UpBlock2D",
+    "CrossAttnUpBlock2D",
+    "CrossAttnUpBlock2D",
+    "CrossAttnUpBlock2D"
+  ],
+  "upcast_attention": null,
+  "use_linear_projection": false
+}

unet/diffusion_pytorch_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2f5e504080331f87d627213f5bde1f0ef8fc44571f884ee4ab9d67aabced37ff
+size 1719125304

vae/config.json

@@ -0,0 +1,38 @@
+{
+  "_class_name": "AutoencoderKL",
+  "_diffusers_version": "0.41.0.dev0",
+  "_name_or_path": "/home/yunustas/InkNoir-Core/hf_cache/hub/models--Lykon--dreamshaper-8/snapshots/a7e52b98680b1ba8ff7bce97c7f9f2e2e5337917/vae",
+  "act_fn": "silu",
+  "block_out_channels": [
+    128,
+    256,
+    512,
+    512
+  ],
+  "down_block_types": [
+    "DownEncoderBlock2D",
+    "DownEncoderBlock2D",
+    "DownEncoderBlock2D",
+    "DownEncoderBlock2D"
+  ],
+  "force_upcast": true,
+  "in_channels": 3,
+  "latent_channels": 4,
+  "latents_mean": null,
+  "latents_std": null,
+  "layers_per_block": 2,
+  "mid_block_add_attention": true,
+  "norm_num_groups": 32,
+  "out_channels": 3,
+  "sample_size": 512,
+  "scaling_factor": 0.18215,
+  "shift_factor": null,
+  "up_block_types": [
+    "UpDecoderBlock2D",
+    "UpDecoderBlock2D",
+    "UpDecoderBlock2D",
+    "UpDecoderBlock2D"
+  ],
+  "use_post_quant_conv": true,
+  "use_quant_conv": true
+}

vae/diffusion_pytorch_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ff38c7ec66a4ae76f466a3ade876522bff4fea8add1981cc433f518666cea33d
+size 167335342