Create tiny Ernie-Image modular pipeline for CI tests

model_index.json

@@ -0,0 +1,24 @@
+{
+  "_class_name": "ErnieImagePipeline",
+  "_diffusers_version": "0.38.0.dev0",
+  "scheduler": [
+    "diffusers",
+    "FlowMatchEulerDiscreteScheduler"
+  ],
+  "text_encoder": [
+    "transformers",
+    "GPT2Model"
+  ],
+  "tokenizer": [
+    "transformers",
+    "GPT2TokenizerFast"
+  ],
+  "transformer": [
+    "diffusers",
+    "ErnieImageTransformer2DModel"
+  ],
+  "vae": [
+    "diffusers",
+    "AutoencoderKLFlux2"
+  ]
+}

modular_model_index.json

@@ -0,0 +1,75 @@
+{
+  "_blocks_class_name": "ErnieImageAutoBlocks",
+  "_class_name": "ErnieImageModularPipeline",
+  "_diffusers_version": "0.38.0.dev0",
+  "scheduler": [
+    null,
+    null,
+    {
+      "pretrained_model_name_or_path": "akshan-main/tiny-ernie-image-modular-pipe",
+      "subfolder": "scheduler",
+      "type_hint": [
+        "diffusers",
+        "FlowMatchEulerDiscreteScheduler"
+      ],
+      "revision": null,
+      "variant": null
+    }
+  ],
+  "text_encoder": [
+    null,
+    null,
+    {
+      "pretrained_model_name_or_path": "akshan-main/tiny-ernie-image-modular-pipe",
+      "subfolder": "text_encoder",
+      "type_hint": [
+        "transformers",
+        "GPT2Model"
+      ],
+      "revision": null,
+      "variant": null
+    }
+  ],
+  "tokenizer": [
+    null,
+    null,
+    {
+      "pretrained_model_name_or_path": "akshan-main/tiny-ernie-image-modular-pipe",
+      "subfolder": "tokenizer",
+      "type_hint": [
+        "transformers",
+        "GPT2TokenizerFast"
+      ],
+      "revision": null,
+      "variant": null
+    }
+  ],
+  "transformer": [
+    null,
+    null,
+    {
+      "pretrained_model_name_or_path": "akshan-main/tiny-ernie-image-modular-pipe",
+      "subfolder": "transformer",
+      "type_hint": [
+        "diffusers",
+        "ErnieImageTransformer2DModel"
+      ],
+      "revision": null,
+      "variant": null
+    }
+  ],
+  "vae": [
+    null,
+    null,
+    {
+      "pretrained_model_name_or_path": "akshan-main/tiny-ernie-image-modular-pipe",
+      "subfolder": "vae",
+      "type_hint": [
+        "diffusers",
+        "AutoencoderKLFlux2"
+      ],
+      "revision": null,
+      "variant": null
+    }
+  ]
+}

scheduler/scheduler_config.json

@@ -0,0 +1,18 @@
+{
+  "_class_name": "FlowMatchEulerDiscreteScheduler",
+  "_diffusers_version": "0.38.0.dev0",
+  "base_image_seq_len": 256,
+  "base_shift": 0.5,
+  "invert_sigmas": false,
+  "max_image_seq_len": 4096,
+  "max_shift": 1.15,
+  "num_train_timesteps": 1000,
+  "shift": 1.0,
+  "shift_terminal": null,
+  "stochastic_sampling": false,
+  "time_shift_type": "exponential",
+  "use_beta_sigmas": false,
+  "use_dynamic_shifting": false,
+  "use_exponential_sigmas": false,
+  "use_karras_sigmas": false
+}

text_encoder/config.json

@@ -0,0 +1,34 @@
+{
+  "activation_function": "gelu_new",
+  "add_cross_attention": false,
+  "architectures": [
+    "GPT2Model"
+  ],
+  "attn_pdrop": 0.1,
+  "bos_token_id": 50256,
+  "dtype": "float32",
+  "embd_pdrop": 0.1,
+  "eos_token_id": 50256,
+  "initializer_range": 0.02,
+  "layer_norm_epsilon": 1e-05,
+  "model_type": "gpt2",
+  "n_embd": 32,
+  "n_head": 4,
+  "n_inner": null,
+  "n_layer": 2,
+  "n_positions": 128,
+  "pad_token_id": null,
+  "reorder_and_upcast_attn": false,
+  "resid_pdrop": 0.1,
+  "scale_attn_by_inverse_layer_idx": false,
+  "scale_attn_weights": true,
+  "summary_activation": null,
+  "summary_first_dropout": 0.1,
+  "summary_proj_to_labels": true,
+  "summary_type": "cls_index",
+  "summary_use_proj": true,
+  "tie_word_embeddings": true,
+  "transformers_version": "5.5.3",
+  "use_cache": true,
+  "vocab_size": 1000
+}

text_encoder/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f62093e4aaefc16cade077e5db24467f376a30bea8f1feeb800ce0211731c370
+size 248536

tokenizer/tokenizer.json

@@ -0,0 +1,173 @@
+{
+  "version": "1.0",
+  "truncation": null,
+  "padding": null,
+  "added_tokens": [
+    {
+      "id": 0,
+      "content": "<|endoftext|>",
+      "single_word": false,
+      "lstrip": false,
+      "rstrip": false,
+      "normalized": false,
+      "special": true
+    }
+  ],
+  "normalizer": null,
+  "pre_tokenizer": {
+    "type": "ByteLevel",
+    "add_prefix_space": false,
+    "trim_offsets": true,
+    "use_regex": true
+  },
+  "post_processor": {
+    "type": "TemplateProcessing",
+    "single": [
+      {
+        "Sequence": {
+          "id": "A",
+          "type_id": 0
+        }
+      }
+    ],
+    "pair": [
+      {
+        "Sequence": {
+          "id": "A",
+          "type_id": 0
+        }
+      },
+      {
+        "Sequence": {
+          "id": "B",
+          "type_id": 1
+        }
+      }
+    ],
+    "special_tokens": {}
+  },
+  "decoder": {
+    "type": "ByteLevel",
+    "add_prefix_space": true,
+    "trim_offsets": true,
+    "use_regex": true
+  },
+  "model": {
+    "type": "BPE",
+    "dropout": null,
+    "unk_token": "<|endoftext|>",
+    "continuing_subword_prefix": null,
+    "end_of_word_suffix": null,
+    "fuse_unk": false,
+    "byte_fallback": false,
+    "ignore_merges": false,
+    "vocab": {
+      "<|endoftext|>": 0,
+      "H": 1,
+      "a": 2,
+      "d": 3,
+      "e": 4,
+      "h": 5,
+      "i": 6,
+      "l": 7,
+      "o": 8,
+      "r": 9,
+      "s": 10,
+      "t": 11,
+      "w": 12,
+      "Ġ": 13,
+      "is": 14,
+      "Ġt": 15,
+      "He": 16,
+      "es": 17,
+      "his": 18,
+      "ld": 19,
+      "ll": 20,
+      "or": 21,
+      "wor": 22,
+      "Ġa": 23,
+      "Ġis": 24,
+      "Ġwor": 25,
+      "Ġtes": 26,
+      "Ġthis": 27,
+      "Hell": 28,
+      "Ġworld": 29,
+      "Ġtest": 30,
+      "Hello": 31
+    },
+    "merges": [
+      [
+        "i",
+        "s"
+      ],
+      [
+        "Ġ",
+        "t"
+      ],
+      [
+        "H",
+        "e"
+      ],
+      [
+        "e",
+        "s"
+      ],
+      [
+        "h",
+        "is"
+      ],
+      [
+        "l",
+        "d"
+      ],
+      [
+        "l",
+        "l"
+      ],
+      [
+        "o",
+        "r"
+      ],
+      [
+        "w",
+        "or"
+      ],
+      [
+        "Ġ",
+        "a"
+      ],
+      [
+        "Ġ",
+        "is"
+      ],
+      [
+        "Ġ",
+        "wor"
+      ],
+      [
+        "Ġt",
+        "es"
+      ],
+      [
+        "Ġt",
+        "his"
+      ],
+      [
+        "He",
+        "ll"
+      ],
+      [
+        "Ġwor",
+        "ld"
+      ],
+      [
+        "Ġtes",
+        "t"
+      ],
+      [
+        "Hell",
+        "o"
+      ]
+    ]
+  }
+}

tokenizer/tokenizer_config.json

@@ -0,0 +1,8 @@
+{
+  "backend": "tokenizers",
+  "bos_token": "<|endoftext|>",
+  "eos_token": "<|endoftext|>",
+  "model_max_length": 1000000000000000019884624838656,
+  "pad_token": "<|endoftext|>",
+  "tokenizer_class": "GPT2TokenizerFast"
+}

transformer/config.json

@@ -0,0 +1,20 @@
+{
+  "_class_name": "ErnieImageTransformer2DModel",
+  "_diffusers_version": "0.38.0.dev0",
+  "eps": 1e-06,
+  "ffn_hidden_size": 64,
+  "hidden_size": 32,
+  "in_channels": 16,
+  "num_attention_heads": 4,
+  "num_layers": 1,
+  "out_channels": 16,
+  "patch_size": 1,
+  "qk_layernorm": true,
+  "rope_axes_dim": [
+    1,
+    1,
+    2
+  ],
+  "rope_theta": 256,
+  "text_in_dim": 32
+}

transformer/diffusion_pytorch_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:24bb88cb10d4e3644fcef8a04a9aa1fb06591b8f631640766716c9d7cc58be83
+size 89936

vae/config.json

@@ -0,0 +1,40 @@
+{
+  "_class_name": "AutoencoderKLFlux2",
+  "_diffusers_version": "0.38.0.dev0",
+  "act_fn": "silu",
+  "batch_norm_eps": 0.0001,
+  "batch_norm_momentum": 0.1,
+  "block_out_channels": [
+    16,
+    32,
+    32,
+    32
+  ],
+  "decoder_block_out_channels": null,
+  "down_block_types": [
+    "DownEncoderBlock2D",
+    "DownEncoderBlock2D",
+    "DownEncoderBlock2D",
+    "DownEncoderBlock2D"
+  ],
+  "force_upcast": true,
+  "in_channels": 3,
+  "latent_channels": 4,
+  "layers_per_block": 1,
+  "mid_block_add_attention": true,
+  "norm_num_groups": 4,
+  "out_channels": 3,
+  "patch_size": [
+    2,
+    2
+  ],
+  "sample_size": 64,
+  "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 sha256:01ccf2698551093e5d2608d75e22e7a8dba6a23ae4118434d33ca0382c485040
+size 1286060