add flash-attn and install glm4-voice

pyproject.toml

@@ -12,26 +12,33 @@ license     = {text = "MIT"}
 authors     = [{name = "MoonshotAI", email = "contact@moonshot.ai"}]
 
 dependencies = [
-  "torch==2.4.1",
-  "torchaudio==2.4.1",
-  "packaging",
-  "jinja2",
+    # Core ML dependencies
+    "torch>=2.0.0",
+    "torchaudio>=2.0.0",
+    "flash-attn>=2.0.0",
+    "transformers",
+    "accelerate",
+    "diffusers",
+    "conformer==0.3.2",
+    "torchdyn==1.0.6",
+    "librosa",
+    "soundfile",
+    "sox",
     "openai-whisper",
-  "jsonlines",
     "pandas",
+    "jsonlines",
+    "datasets",
+    "huggingface_hub",
+    "decord",
+    "blobfile",
+    "packaging",
+    "jinja2",
     "validators",
     "sty",
-  "transformers",
-  "librosa",
-  "accelerate",
-  "aiohttp",
     "colorama",
     "omegaconf==2.3.0",
-  "sox",
-  "six==1.16.0",
+    "six>=1.16.0",
     "hyperpyyaml",
-  "conformer==0.3.2",
-  "diffusers",
     "pillow",
     "sentencepiece",
     "easydict",
@@ -42,15 +49,11 @@ dependencies = [
     "rich",
     "wget",
     "gdown",
-  "datasets",
-  "torchdyn==1.0.6",
-  "huggingface_hub",
-  "loguru",
-  "blobfile",
     "timm",
     "sacrebleu==1.5.1",
-  "soundfile",
-  "tqdm"
+    "tqdm",
+    "loguru",
+    "flash-attn"
 ]
 
 [tool.setuptools]
@@ -60,5 +63,8 @@ include-package-data = true
 where = ["."]
 include = ["kimia_infer*"]
 
+[tool.setuptools.package-data]
+"kimia_infer" = ["**/*"]
+
 [project.urls]
 Repository = "https://github.com/MoonshotAI/Kimi-Audio"