Oct 18 21:53 Add files using upload-large-folder tool 21:52 Add files using upload-large-folder tool 21:18 Add files using upload-large-folder tool 21:17 Add files using upload-large-folder tool 20:37 Add files using upload-large-folder tool 20:37 Add files using upload-large-folder tool 19:56 Add files using upload-large-folder tool 19:55 Add files using upload-large-folder tool 19:55 Add files using upload-large-folder tool 19:55 Add files using upload-large-folder tool 19:55 Add files using upload-large-folder tool 19:55 Add files using upload-large-folder tool 04:53 Model card: experts quantized from the FP8 release are the default (full model and REAP-50) 04:52 Add GLM-5.3-Flash-Q23-FP8-assembled.gguf Oct 17 19:56 Add reap50/reap50-expert-map.txt 19:55 Add reap50/GLM-5.3-Flash-REAP50-Q23-FP8-assembled.gguf 16:17 Add experts-q23-fp8.gguf.quality.json 16:17 Add experts-q23-fp8.gguf Oct 15 17:31 Model card: current measurements (one and two GPUs, RTX 3070), HumanEval results, profiles, context length Oct 13 10:58 Document REAP-224 and recommend full Q23 for output quality Record full HumanEval results with build/protocol caveats and the unmet REAP-224 release gate. List the existing REAP-224 weights, checksum and metadata. Model card only; weight files are unchanged. Oct 12 00:28 Add reap224/GLM-5.3-Flash-Q23-REAP224-assembled.gguf 00:20 Add reap224/coverage-keep224.json 00:19 Add reap224/plan-keep224.report.json 00:17 Add reap224/plan-keep224.txt Oct 10 20:51 Add GLM-5.3-Flash-Q23-assembled.gguf 20:40 Add calibration/coverage-q23-tr16.json 20:38 Add experts-q23.gguf.quality.json 20:36 Add experts-q22.gguf.quality.json 20:36 Add experts-q22.gguf 19:35 Add README.md 19:35 Add reap50/GLM-5.3-Flash-REAP50-Q23-assembled.gguf 10:50 Add README.md 10:49 Add reap50/reap50-coverage.json 10:47 Add reap50/reap50-q23-assembled.provenance.json Oct 8 09:23 Upload Q2Q3 expert sidecar Oct 7 06:07 Document expert conversion, reproduction, licensing and attribution 05:31 initial commit