diffsynth-engine 0.6.1.dev33__py3-none-any.whl → 0.6.1.dev35__py3-none-any.whl
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- diffsynth_engine/__init__.py +4 -0
- diffsynth_engine/conf/models/z_image/qwen3_config.json +30 -0
- diffsynth_engine/conf/tokenizers/z_image/tokenizer/merges.txt +151388 -0
- diffsynth_engine/conf/tokenizers/z_image/tokenizer/tokenizer.json +757480 -0
- diffsynth_engine/conf/tokenizers/z_image/tokenizer/tokenizer_config.json +239 -0
- diffsynth_engine/conf/tokenizers/z_image/tokenizer/vocab.json +1 -0
- diffsynth_engine/configs/__init__.py +4 -0
- diffsynth_engine/configs/pipeline.py +44 -1
- diffsynth_engine/models/basic/attention.py +2 -2
- diffsynth_engine/models/qwen_image/qwen_image_dit.py +5 -5
- diffsynth_engine/models/qwen_image/qwen_image_vae.py +0 -1
- diffsynth_engine/models/z_image/__init__.py +11 -0
- diffsynth_engine/models/z_image/qwen3.py +124 -0
- diffsynth_engine/models/z_image/z_image_dit.py +602 -0
- diffsynth_engine/pipelines/__init__.py +2 -0
- diffsynth_engine/pipelines/qwen_image.py +4 -3
- diffsynth_engine/pipelines/z_image.py +377 -0
- diffsynth_engine/utils/constants.py +3 -0
- diffsynth_engine/utils/process_group.py +1 -1
- {diffsynth_engine-0.6.1.dev33.dist-info → diffsynth_engine-0.6.1.dev35.dist-info}/METADATA +1 -1
- {diffsynth_engine-0.6.1.dev33.dist-info → diffsynth_engine-0.6.1.dev35.dist-info}/RECORD +24 -15
- {diffsynth_engine-0.6.1.dev33.dist-info → diffsynth_engine-0.6.1.dev35.dist-info}/WHEEL +0 -0
- {diffsynth_engine-0.6.1.dev33.dist-info → diffsynth_engine-0.6.1.dev35.dist-info}/licenses/LICENSE +0 -0
- {diffsynth_engine-0.6.1.dev33.dist-info → diffsynth_engine-0.6.1.dev35.dist-info}/top_level.txt +0 -0
diffsynth_engine/__init__.py
CHANGED
|
@@ -6,11 +6,13 @@ from .configs import (
|
|
|
6
6
|
WanSpeech2VideoPipelineConfig,
|
|
7
7
|
QwenImagePipelineConfig,
|
|
8
8
|
HunyuanPipelineConfig,
|
|
9
|
+
ZImagePipelineConfig,
|
|
9
10
|
SDStateDicts,
|
|
10
11
|
SDXLStateDicts,
|
|
11
12
|
FluxStateDicts,
|
|
12
13
|
WanStateDicts,
|
|
13
14
|
QwenImageStateDicts,
|
|
15
|
+
ZImageStateDicts,
|
|
14
16
|
AttnImpl,
|
|
15
17
|
SpargeAttentionParams,
|
|
16
18
|
VideoSparseAttentionParams,
|
|
@@ -55,11 +57,13 @@ __all__ = [
|
|
|
55
57
|
"WanSpeech2VideoPipelineConfig",
|
|
56
58
|
"QwenImagePipelineConfig",
|
|
57
59
|
"HunyuanPipelineConfig",
|
|
60
|
+
"ZImagePipelineConfig",
|
|
58
61
|
"SDStateDicts",
|
|
59
62
|
"SDXLStateDicts",
|
|
60
63
|
"FluxStateDicts",
|
|
61
64
|
"WanStateDicts",
|
|
62
65
|
"QwenImageStateDicts",
|
|
66
|
+
"ZImageStateDicts",
|
|
63
67
|
"AttnImpl",
|
|
64
68
|
"SpargeAttentionParams",
|
|
65
69
|
"VideoSparseAttentionParams",
|
|
@@ -0,0 +1,30 @@
|
|
|
1
|
+
{
|
|
2
|
+
"architectures": [
|
|
3
|
+
"Qwen3ForCausalLM"
|
|
4
|
+
],
|
|
5
|
+
"attention_bias": false,
|
|
6
|
+
"attention_dropout": 0.0,
|
|
7
|
+
"bos_token_id": 151643,
|
|
8
|
+
"eos_token_id": 151645,
|
|
9
|
+
"head_dim": 128,
|
|
10
|
+
"hidden_act": "silu",
|
|
11
|
+
"hidden_size": 2560,
|
|
12
|
+
"initializer_range": 0.02,
|
|
13
|
+
"intermediate_size": 9728,
|
|
14
|
+
"max_position_embeddings": 40960,
|
|
15
|
+
"max_window_layers": 36,
|
|
16
|
+
"model_type": "qwen3",
|
|
17
|
+
"num_attention_heads": 32,
|
|
18
|
+
"num_hidden_layers": 36,
|
|
19
|
+
"num_key_value_heads": 8,
|
|
20
|
+
"rms_norm_eps": 1e-06,
|
|
21
|
+
"rope_scaling": null,
|
|
22
|
+
"rope_theta": 1000000,
|
|
23
|
+
"sliding_window": null,
|
|
24
|
+
"tie_word_embeddings": true,
|
|
25
|
+
"torch_dtype": "bfloat16",
|
|
26
|
+
"transformers_version": "4.51.0",
|
|
27
|
+
"use_cache": true,
|
|
28
|
+
"use_sliding_window": false,
|
|
29
|
+
"vocab_size": 151936
|
|
30
|
+
}
|