21 October 2026 · GTX 1060 6 GB · Packet Foundry

Predator media: CogVideoX Q8 · original bridge, unusable noise

TL;DR

2 retained outputs · 2 accepted results · 2 recorded attempts · median Factory trace 2435.9s Technical delivery and media decoding do not establish visual or musical quality. Timings include adapter overhead; cache behavior differs between runs. Failed and cancelled attempts remain in the evidence.

2 retained outputs · 2 accepted results · 2 recorded attempts · median Factory trace 2435.9s

Technical delivery and media decoding do not establish visual or musical quality. Timings include adapter overhead; cache behavior differs between runs. Failed and cancelled attempts remain in the evidence.

← Complete campaign · Read the article · Download evidence

cogvideox2b-q8-subquad-text-v2 · succeeded

CogVideoX 2B local Q8_0 · seed 42

Seed 42 · 2484.046 s Factory trace

Download MP4 · 1.44 MiB

  • All five reviewed frames are colored noise.
  • No recognizable fox, snow clearing or requested scene appears.
  • Successful container decoding and trace verification do not establish a working semantic generation integration.
  • Unusable colored-noise output from the original custom bridge. Retained as failure evidence; bridge positional/sampler mismatch later corrected.
Exact prompt
A realistic red fox walks steadily across a snowy clearing from the left side of the frame to the right. Show the fox in full body from the side. Its four legs visibly alternate with each walking step, its body advances across the ground, and its bushy tail swings gently. Near the end the fox stops walking and turns its head toward the camera. The camera stays completely fixed in one wide shot. Continuous visible animal movement throughout the walk. Natural daylight, realistic anatomy, one continuous six-second shot, no cuts, no text.
Parameters and prompt details
{
  "parameters": {
    "2 CLIPLoaderGGUF": {
      "clip_name": "t5-v1_1-xxl-encoder-Q4_K_M.gguf",
      "type": "cogvideox"
    },
    "3 VAELoader": {
      "vae_name": "cogvideox-2b-vae.safetensors"
    },
    "7 EmptyHunyuanLatentVideo": {
      "width": 512,
      "height": 320,
      "length": 49,
      "batch_size": 1
    },
    "8 PFCogWholeStageSubQuadSampler": {
      "seed": 42,
      "steps": 30,
      "cfg": 6.0,
      "sampler_name": "ddim",
      "scheduler": "normal",
      "denoise": 1.0
    },
    "90 PFWholeStageDecode": {
      "tile_size": 256,
      "overlap": 64,
      "temporal_size": 32,
      "temporal_overlap": 8
    },
    "91 CreateVideo": {
      "fps": 8.0
    },
    "92 SaveVideo": {
      "format": "mp4",
      "format.codec": "h264"
    }
  },
  "prompt_details": {
    "conditioning_texts": {
      "4.positive": "Create one video clip from the supplied scene description.\n\nSource generation_brief:\nA realistic red fox walks steadily across a snowy clearing from the left side of the frame to the right. Show the fox in full body from the side. Its four legs visibly alternate with each walking step, its body advances across the ground, and its bushy tail swings gently. Near the end the fox stops walking and turns its head toward the camera. The camera stays completely fixed in one wide shot. Continuous visible animal movement throughout the walk. Natural daylight, realistic anatomy, one continuous six-second shot, no cuts, no text.",
      "4.negative": "motionless animal, frozen pose, static image, slideshow, distorted anatomy, blurry, flickering, subtitles, watermark"
    },
    "actual_positive_text": "Create one video clip from the supplied scene description.\n\nSource generation_brief:\nA realistic red fox walks steadily across a snowy clearing from the left side of the frame to the right. Show the fox in full body from the side. Its four legs visibly alternate with each walking step, its body advances across the ground, and its bushy tail swings gently. Near the end the fox stops walking and turns its head toward the camera. The camera stays completely fixed in one wide shot. Continuous visible animal movement throughout the walk. Natural daylight, realistic anatomy, one continuous six-second shot, no cuts, no text."
  }
}
Timings, cache and resources
{
  "timing": {
    "factory_trace_seconds": 2484.046425,
    "comfy_server_seconds": 2481.278,
    "comfy_printed_seconds": null,
    "stages_seconds": {
      "text_encoder": 35.95299999997951,
      "generator": 2294.7970000000205,
      "video_decoder": 112.85899999999674
    },
    "limits": "Factory trace and server totals have different boundaries. Observer node intervals include preparation/event delivery; not isolated kernel or weight-loading times."
  },
  "cache": {
    "cached_nodes": [],
    "same_runtime_retained": null,
    "immediate_warm_claim": null,
    "context": {}
  },
  "resources": {
    "sample_count": 1211,
    "scope": "Sampled values; GPU use is device-wide, not continuous per-process physical residency.",
    "minimum_ram_available_bytes": 8916254720,
    "minimum_commit_available_bytes": 16130121728,
    "minimum_system_drive_free_bytes": 20241936384,
    "maximum_gpu_total_used_mib": 5970
  }
}
Verification and checksums
{
  "verification": {
    "full_decode": true,
    "checksum": true,
    "acknowledged": true,
    "trace_attachment": true,
    "native_experiment": true,
    "quality_scope": "Limited sampled review; no systematic full-resolution/per-trait/full-rate quality score.",
    "independent_public_copy_decode": true
  },
  "media": [
    {
      "path": "media/cog-01.mp4",
      "mime_type": "video/mp4",
      "size_bytes": 1505622,
      "sha256": "4dd3b927cac4950eccd9fb7fbe1df583f7ec34d099c812da84add0cb6bc2012c",
      "source_sha256": "a017a87ace7a579134323b4bcd1f0fd6418556a9ce8fb0a8d1259119f651a678",
      "metadata_handling": "Container metadata stripped with stream copy; encoded audio/video streams not re-encoded.",
      "width": 512,
      "height": 320,
      "duration_seconds": 6.125,
      "frames": 49,
      "fps": 8.0,
      "public_copy_full_decode_passed": true,
      "source_content_preserved": "Encodedstream hashes identical aftermetadata-only remux"
    }
  ]
}

Workflow snapshot

Download this run record

cogvideox2b-q8-subquad-text-v2 · succeeded

CogVideoX 2B local Q8_0 · seed 43

Seed 43 · 2387.768 s Factory trace

Download MP4 · 1.45 MiB

  • All five reviewed frames are colored noise.
  • No recognizable fox, snow clearing or requested scene appears.
  • Native video/container decoding, relay, trace attachments and comparable experiment export succeeded; semantic adapter integration failed.
  • Unusable colored-noise output from the original custom bridge. Retained as failure evidence; bridge positional/sampler mismatch later corrected.
Exact prompt
A realistic red fox walks steadily across a snowy clearing from the left side of the frame to the right. Show the fox in full body from the side. Its four legs visibly alternate with each walking step, its body advances across the ground, and its bushy tail swings gently. Near the end the fox stops walking and turns its head toward the camera. The camera stays completely fixed in one wide shot. Continuous visible animal movement throughout the walk. Natural daylight, realistic anatomy, one continuous six-second shot, no cuts, no text.
Parameters and prompt details
{
  "parameters": {
    "2 CLIPLoaderGGUF": {
      "clip_name": "t5-v1_1-xxl-encoder-Q4_K_M.gguf",
      "type": "cogvideox"
    },
    "3 VAELoader": {
      "vae_name": "cogvideox-2b-vae.safetensors"
    },
    "7 EmptyHunyuanLatentVideo": {
      "width": 512,
      "height": 320,
      "length": 49,
      "batch_size": 1
    },
    "8 PFCogWholeStageSubQuadSampler": {
      "seed": 43,
      "steps": 30,
      "cfg": 6.0,
      "sampler_name": "ddim",
      "scheduler": "normal",
      "denoise": 1.0
    },
    "90 PFWholeStageDecode": {
      "tile_size": 256,
      "overlap": 64,
      "temporal_size": 32,
      "temporal_overlap": 8
    },
    "91 CreateVideo": {
      "fps": 8.0
    },
    "92 SaveVideo": {
      "format": "mp4",
      "format.codec": "h264"
    }
  },
  "prompt_details": {
    "conditioning_texts": {
      "4.positive": "Create one video clip from the supplied scene description.\n\nSource generation_brief:\nA realistic red fox walks steadily across a snowy clearing from the left side of the frame to the right. Show the fox in full body from the side. Its four legs visibly alternate with each walking step, its body advances across the ground, and its bushy tail swings gently. Near the end the fox stops walking and turns its head toward the camera. The camera stays completely fixed in one wide shot. Continuous visible animal movement throughout the walk. Natural daylight, realistic anatomy, one continuous six-second shot, no cuts, no text.",
      "4.negative": "motionless animal, frozen pose, static image, slideshow, distorted anatomy, blurry, flickering, subtitles, watermark"
    },
    "actual_positive_text": "Create one video clip from the supplied scene description.\n\nSource generation_brief:\nA realistic red fox walks steadily across a snowy clearing from the left side of the frame to the right. Show the fox in full body from the side. Its four legs visibly alternate with each walking step, its body advances across the ground, and its bushy tail swings gently. Near the end the fox stops walking and turns its head toward the camera. The camera stays completely fixed in one wide shot. Continuous visible animal movement throughout the walk. Natural daylight, realistic anatomy, one continuous six-second shot, no cuts, no text."
  }
}
Timings, cache and resources
{
  "timing": {
    "factory_trace_seconds": 2387.768021,
    "comfy_server_seconds": 2383.053,
    "comfy_printed_seconds": null,
    "stages_seconds": {
      "generator": 2269.5630000000237,
      "video_decoder": 109.92199999996228
    },
    "limits": "Factory trace and server totals have different boundaries. Observer node intervals include preparation/event delivery; not isolated kernel or weight-loading times."
  },
  "cache": {
    "cached_nodes": [
      "2",
      "4",
      "1",
      "3",
      "7"
    ],
    "same_runtime_retained": true,
    "immediate_warm_claim": null,
    "context": {}
  },
  "resources": {
    "sample_count": 1164,
    "scope": "Sampled values; GPU use is device-wide, not continuous per-process physical residency.",
    "minimum_ram_available_bytes": 10545700864,
    "minimum_commit_available_bytes": 15607156736,
    "minimum_system_drive_free_bytes": 20227006464,
    "maximum_gpu_total_used_mib": 5895
  }
}
Verification and checksums
{
  "verification": {
    "full_decode": true,
    "checksum": true,
    "acknowledged": true,
    "trace_attachment": true,
    "native_experiment": true,
    "quality_scope": "Limited sampled review; no systematic full-resolution/per-trait/full-rate quality score.",
    "independent_public_copy_decode": true
  },
  "media": [
    {
      "path": "media/cog-02.mp4",
      "mime_type": "video/mp4",
      "size_bytes": 1518836,
      "sha256": "32b9da79aadb9a01b0fa08091ec65fd2d369a5b488524780ab7829c0bc50e7df",
      "source_sha256": "a21a9a6ca5fa748777d5907cdb77a0a4f7bc655aae69bd7b36bd8a82a1c07490",
      "metadata_handling": "Container metadata stripped with stream copy; encoded audio/video streams not re-encoded.",
      "width": 512,
      "height": 320,
      "duration_seconds": 6.125,
      "frames": 49,
      "fps": 8.0,
      "public_copy_full_decode_passed": true,
      "source_content_preserved": "Encodedstream hashes identical aftermetadata-only remux"
    }
  ]
}

Workflow snapshot

Download this run record