Added model type environment variable so we can actually attempt to benchmark with the right payload.

Co-authored-by: Edgar Lin <edgarlin2000@gmail.com>

lib/backend.py

@@ -286,7 +286,7 @@ class Backend:
         message = {
             key: value
             for (key, value) in (dataclasses.asdict(auth_data).items())
-            if key != "signature"
+            if key != "signature" and key != "__request_id"
         }
         if auth_data.reqnum < (self.reqnum - MSG_HISTORY_LEN):
             log.debug(
@@ -296,7 +296,7 @@ class Backend:
         elif message in self.msg_history:
             log.debug(f"message: {message} already in message history")
             return False
-        elif verify_signature(json.dumps(message, indent=4), auth_data.signature):
+        elif verify_signature(json.dumps(message, indent=4, sort_keys=True), auth_data.signature):
             self.reqnum = max(auth_data.reqnum, self.reqnum)
             self.msg_history.append(message)
             self.msg_history = self.msg_history[-MSG_HISTORY_LEN:]

lib/data_types.py

@@ -65,12 +65,12 @@ class ApiPayload(ABC):
 class AuthData:
     """data used to authenticate requester"""
 
-    signature: str
     cost: str
     endpoint: str
     reqnum: int
-    url: str
     request_idx: int
+    signature: str
+    url: str
 
     @classmethod
     def from_json_msg(cls, json_msg: Dict[str, Any]):

lib/metrics.py

@@ -180,6 +180,10 @@ class Metrics:
             return  # nothing to do
 
         for report_addr in self.report_addr:
+            # TODO: Add a Redis subscriber queue for delete_requests
+            if report_addr == "https://cloud.vast.ai/api/v0":
+                # Patch: ignore the Redis API report_addr
+                continue
             sent_success = True
             sent_failed  = True
 

start_server.sh

@@ -9,9 +9,10 @@ ENV_PATH="$WORKSPACE_DIR/worker-env"
 DEBUG_LOG="$WORKSPACE_DIR/debug.log"
 PYWORKER_LOG="$WORKSPACE_DIR/pyworker.log"
 
-REPORT_ADDR="${REPORT_ADDR:-https://cloud.vast.ai/api/v0,https://run.vast.ai}"
+REPORT_ADDR="${REPORT_ADDR:-https://run.vast.ai}"
 USE_SSL="${USE_SSL:-true}"
 WORKER_PORT="${WORKER_PORT:-3000}"
+MODEL_TYPE="${MODEL_TYPE:-image}"
 mkdir -p "$WORKSPACE_DIR"
 cd "$WORKSPACE_DIR"
 

workers/comfyui-json/data_types.py

@@ -1,3 +1,5 @@
+import dataclasses
+from enum import Enum
 import os
 import sys
 import random
@@ -13,6 +15,12 @@ from lib.data_types import ApiPayload, JsonDataException
 
 log = logging.getLogger(__file__)
 
+class ModelType(Enum):
+    image = "image"
+    audio = "audio"
+    video = "video"
+
+
 def count_workload() -> float:
     # Always 100.0 where there is a single instance of ComfyUI handling requests
     # Results will indicate % or a job completed per second.  Avoids sub 0.1 sec performance indication
@@ -21,6 +29,11 @@ def count_workload() -> float:
 @dataclasses.dataclass
 class ComfyWorkflowData(ApiPayload):
     input: dict
+    model_type: ModelType = dataclasses.field(
+        default_factory=lambda: ModelType(
+            os.environ.get("MODEL_TYPE", "image").lower()
+        )
+    )
 
     @classmethod
     def for_test(cls):
@@ -30,15 +43,17 @@ class ComfyWorkflowData(ApiPayload):
         Example: SD1.5, simple image gen 10000 steps, 512px x 512px will run for approximately 9 minutes @ ~18 it/s (RTX 4090)
         """
         # Try to load benchmark.json
-        benchmark_file = Path("workers/comfyui-json/misc/benchmark.json")
-        
+        #Note:  We should cross check with Rob if the audio sample benchmark file is correct
+        model_type = ModelType(os.environ.get("MODEL_TYPE", "image").lower())
+        benchmark_file = Path(f"workers/comfyui-json/misc/benchmark_{model_type.value}.json")
         if benchmark_file.exists():
             try:
                 with open(benchmark_file, "r") as f:
                     benchmark_workflow = json.load(f)
+                log.info(f"using benchmark json file for {model_type.value}")
                 return cls(
                     input={
-                        "request_id": f"test-{random.randint(1000, 99999)}",
+                        "request_id": f"{model_type.value}-{random.randint(1000, 99999)}",
                         "workflow_json": benchmark_workflow
                     }
                 )

workers/comfyui-json/misc/benchmark_audio.json

@@ -0,0 +1,118 @@
+{
+        "3": {
+            "inputs": {
+            "seed": 98942092715729,
+            "steps": 50,
+            "cfg": 4.98,
+            "sampler_name": "dpmpp_3m_sde_gpu",
+            "scheduler": "exponential",
+            "denoise": 1,
+            "model": [
+                "4",
+                0
+            ],
+            "positive": [
+                "6",
+                0
+            ],
+            "negative": [
+                "7",
+                0
+            ],
+            "latent_image": [
+                "11",
+                0
+            ]
+            },
+            "class_type": "KSampler",
+            "_meta": {
+            "title": "KSampler"
+            }
+        },
+        "4": {
+            "inputs": {
+            "ckpt_name": "stable-audio-open-1.0.safetensors"
+            },
+            "class_type": "CheckpointLoaderSimple",
+            "_meta": {
+            "title": "Load Checkpoint"
+            }
+        },
+        "6": {
+            "inputs": {
+            "text": "heaven church electronic dance music",
+            "clip": [
+                "10",
+                0
+            ]
+            },
+            "class_type": "CLIPTextEncode",
+            "_meta": {
+            "title": "CLIP Text Encode (Prompt)"
+            }
+        },
+        "7": {
+            "inputs": {
+            "text": "",
+            "clip": [
+                "10",
+                0
+            ]
+            },
+            "class_type": "CLIPTextEncode",
+            "_meta": {
+            "title": "CLIP Text Encode (Prompt)"
+            }
+        },
+        "10": {
+            "inputs": {
+            "clip_name": "t5-base.safetensors",
+            "type": "stable_audio",
+            "device": "default"
+            },
+            "class_type": "CLIPLoader",
+            "_meta": {
+            "title": "Load CLIP"
+            }
+        },
+        "11": {
+            "inputs": {
+            "seconds": 47.6,
+            "batch_size": 1
+            },
+            "class_type": "EmptyLatentAudio",
+            "_meta": {
+            "title": "EmptyLatentAudio"
+            }
+        },
+        "12": {
+            "inputs": {
+            "samples": [
+                "3",
+                0
+            ],
+            "vae": [
+                "4",
+                2
+            ]
+            },
+            "class_type": "VAEDecodeAudio",
+            "_meta": {
+            "title": "VAEDecodeAudio"
+            }
+        },
+        "13": {
+            "inputs": {
+            "filename_prefix": "audio/ComfyUI",
+            "audioUI": "",
+            "audio": [
+                "12",
+                0
+            ]
+            },
+            "class_type": "SaveAudio",
+            "_meta": {
+            "title": "SaveAudio"
+            }
+        }
+    }

workers/comfyui-json/misc/benchmark_image.json

@@ -0,0 +1,107 @@
+{
+      "3": {
+        "inputs": {
+          "seed": 588445435278533,
+          "steps": 20,
+          "cfg": 8,
+          "sampler_name": "euler",
+          "scheduler": "normal",
+          "denoise": 1,
+          "model": [
+            "4",
+            0
+          ],
+          "positive": [
+            "6",
+            0
+          ],
+          "negative": [
+            "7",
+            0
+          ],
+          "latent_image": [
+            "5",
+            0
+          ]
+        },
+        "class_type": "KSampler",
+        "_meta": {
+          "title": "KSampler"
+        }
+      },
+      "4": {
+        "inputs": {
+          "ckpt_name": "v1-5-pruned-emaonly-fp16.safetensors"
+        },
+        "class_type": "CheckpointLoaderSimple",
+        "_meta": {
+          "title": "Load Checkpoint"
+        }
+      },
+      "5": {
+        "inputs": {
+          "width": 512,
+          "height": 512,
+          "batch_size": 1
+        },
+        "class_type": "EmptyLatentImage",
+        "_meta": {
+          "title": "Empty Latent Image"
+        }
+      },
+      "6": {
+        "inputs": {
+          "text": "beautiful scenery nature glass bottle landscape, , purple galaxy bottle,",
+          "clip": [
+            "4",
+            1
+          ]
+        },
+        "class_type": "CLIPTextEncode",
+        "_meta": {
+          "title": "CLIP Text Encode (Prompt)"
+        }
+      },
+      "7": {
+        "inputs": {
+          "text": "text, watermark",
+          "clip": [
+            "4",
+            1
+          ]
+        },
+        "class_type": "CLIPTextEncode",
+        "_meta": {
+          "title": "CLIP Text Encode (Prompt)"
+        }
+      },
+      "8": {
+        "inputs": {
+          "samples": [
+            "3",
+            0
+          ],
+          "vae": [
+            "4",
+            2
+          ]
+        },
+        "class_type": "VAEDecode",
+        "_meta": {
+          "title": "VAE Decode"
+        }
+      },
+      "9": {
+        "inputs": {
+          "filename_prefix": "ComfyUI",
+          "images": [
+            "8",
+            0
+          ]
+        },
+        "class_type": "SaveImage",
+        "_meta": {
+          "title": "Save Image"
+        }
+      }
+    }

workers/comfyui-json/misc/benchmark_video.json

@@ -0,0 +1,216 @@
+{
+    "90": {
+        "inputs": {
+            "clip_name": "umt5_xxl_fp8_e4m3fn_scaled.safetensors",
+            "type": "wan",
+            "device": "default"
+        },
+        "class_type": "CLIPLoader",
+        "_meta": {
+            "title": "Load CLIP"
+        }
+    },
+    "91": {
+        "inputs": {
+            "text": "色调艳丽，过曝，静态，细节模糊不清，字幕，风格，作品，画作，画面，静止，整体发灰，最差质量，低质量，JPEG压缩残留，丑陋的，残缺的，多余的手指，画得不好的手部，画得不好的脸部，畸形的，毁容的，形态畸形的肢体，手指融合，静止不动的画面，杂乱的背景，三条腿，背景人很多，倒着走，裸露，NSFW",
+            "clip": [
+                "90",
+                0
+            ]
+        },
+        "class_type": "CLIPTextEncode",
+        "_meta": {
+            "title": "CLIP Text Encode (Negative Prompt)"
+        }
+    },
+    "92": {
+        "inputs": {
+            "vae_name": "wan_2.1_vae.safetensors"
+        },
+        "class_type": "VAELoader",
+        "_meta": {
+            "title": "Load VAE"
+        }
+    },
+    "93": {
+        "inputs": {
+            "shift": 8.000000000000002,
+            "model": [
+                "101",
+                0
+            ]
+        },
+        "class_type": "ModelSamplingSD3",
+        "_meta": {
+            "title": "ModelSamplingSD3"
+        }
+    },
+    "94": {
+        "inputs": {
+            "shift": 8,
+            "model": [
+                "102",
+                0
+            ]
+        },
+        "class_type": "ModelSamplingSD3",
+        "_meta": {
+            "title": "ModelSamplingSD3"
+        }
+    },
+    "95": {
+        "inputs": {
+            "add_noise": "disable",
+            "noise_seed": 0,
+            "steps": 20,
+            "cfg": 3.5,
+            "sampler_name": "euler",
+            "scheduler": "simple",
+            "start_at_step": 10,
+            "end_at_step": 10000,
+            "return_with_leftover_noise": "disable",
+            "model": [
+                "94",
+                0
+            ],
+            "positive": [
+                "99",
+                0
+            ],
+            "negative": [
+                "91",
+                0
+            ],
+            "latent_image": [
+                "96",
+                0
+            ]
+        },
+        "class_type": "KSamplerAdvanced",
+        "_meta": {
+            "title": "KSampler (Advanced)"
+        }
+    },
+    "96": {
+        "inputs": {
+            "add_noise": "enable",
+            "noise_seed": "__RANDOM_INT__",
+            "steps": 20,
+            "cfg": 3.5,
+            "sampler_name": "euler",
+            "scheduler": "simple",
+            "start_at_step": 0,
+            "end_at_step": 10,
+            "return_with_leftover_noise": "enable",
+            "model": [
+                "93",
+                0
+            ],
+            "positive": [
+                "99",
+                0
+            ],
+            "negative": [
+                "91",
+                0
+            ],
+            "latent_image": [
+                "104",
+                0
+            ]
+        },
+        "class_type": "KSamplerAdvanced",
+        "_meta": {
+            "title": "KSampler (Advanced)"
+        }
+    },
+    "97": {
+        "inputs": {
+            "samples": [
+                "95",
+                0
+            ],
+            "vae": [
+                "92",
+                0
+            ]
+        },
+        "class_type": "VAEDecode",
+        "_meta": {
+            "title": "VAE Decode"
+        }
+    },
+    "98": {
+        "inputs": {
+            "filename_prefix": "video/ComfyUI",
+            "format": "auto",
+            "codec": "auto",
+            "video": [
+                "100",
+                0
+            ]
+        },
+        "class_type": "SaveVideo",
+        "_meta": {
+            "title": "Save Video"
+        }
+    },
+    "99": {
+        "inputs": {
+            "text": "Beautiful young European woman with honey blonde hair gracefully turning her head back over shoulder, gentle smile, bright eyes looking at camera. Hair flowing in slow motion as she turns. Soft natural lighting, clean background, cinematic portrait.",
+            "clip": [
+                "90",
+                0
+            ]
+        },
+        "class_type": "CLIPTextEncode",
+        "_meta": {
+            "title": "CLIP Text Encode (Positive Prompt)"
+        }
+    },
+    "100": {
+        "inputs": {
+            "fps": 16,
+            "images": [
+                "97",
+                0
+            ]
+        },
+        "class_type": "CreateVideo",
+        "_meta": {
+            "title": "Create Video"
+        }
+    },
+    "101": {
+        "inputs": {
+            "unet_name": "wan2.2_t2v_high_noise_14B_fp8_scaled.safetensors",
+            "weight_dtype": "default"
+        },
+        "class_type": "UNETLoader",
+        "_meta": {
+            "title": "Load Diffusion Model"
+        }
+    },
+    "102": {
+        "inputs": {
+            "unet_name": "wan2.2_t2v_low_noise_14B_fp8_scaled.safetensors",
+            "weight_dtype": "default"
+        },
+        "class_type": "UNETLoader",
+        "_meta": {
+            "title": "Load Diffusion Model"
+        }
+    },
+    "104": {
+        "inputs": {
+            "width": 640,
+            "height": 640,
+            "length": 81,
+            "batch_size": 1
+        },
+        "class_type": "EmptyHunyuanLatentVideo",
+        "_meta": {
+            "title": "EmptyHunyuanLatentVideo"
+        }
+    }
+}