{"schemaVersion":"hugging-bay.fit-check.v1","generatedAt":"2026-07-10T04:37:59.514Z","artifactId":"hf-model-davidkim205-rhea-72b-v0-5","repo":"davidkim205/Rhea-72b-v0.5","webUrl":"https://huggingbay.xyz/artifact/hf-model-davidkim205-rhea-72b-v0-5","name":"Rhea-72b-v0.5","paramsB":72,"activeB":72,"quant":"f16","weightFile":null,"fileSizeBytes":1,"ctx":8192,"batch":1,"estimate":{"weightsGb":0,"kvGb":2.68,"overheadGb":0.8,"totalGb":3.48},"gpus":[{"id":"t4-16","label":"NVIDIA T4 16GB (free Colab)","vramGb":16,"unified":false,"verdict":"fits","headroomGb":12.52,"totalGb":3.48},{"id":"l4-24","label":"NVIDIA L4 24GB","vramGb":24,"unified":false,"verdict":"fits","headroomGb":20.52,"totalGb":3.48},{"id":"rtx3060-12","label":"RTX 3060 12GB","vramGb":12,"unified":false,"verdict":"fits","headroomGb":8.52,"totalGb":3.48},{"id":"rtx4080-16","label":"RTX 4080 16GB","vramGb":16,"unified":false,"verdict":"fits","headroomGb":12.52,"totalGb":3.48},{"id":"rtx3090-24","label":"RTX 3090 24GB","vramGb":24,"unified":false,"verdict":"fits","headroomGb":20.52,"totalGb":3.48},{"id":"rtx4090-24","label":"RTX 4090 24GB","vramGb":24,"unified":false,"verdict":"fits","headroomGb":20.52,"totalGb":3.48},{"id":"a100-40","label":"A100 40GB","vramGb":40,"unified":false,"verdict":"fits","headroomGb":36.52,"totalGb":3.48},{"id":"a100-80","label":"A100 80GB","vramGb":80,"unified":false,"verdict":"fits","headroomGb":76.52,"totalGb":3.48},{"id":"h100-80","label":"H100 80GB","vramGb":80,"unified":false,"verdict":"fits","headroomGb":76.52,"totalGb":3.48},{"id":"apple-m-16","label":"Apple M-series 16GB (unified)","vramGb":16,"unified":true,"verdict":"fits","headroomGb":12.52,"totalGb":3.48},{"id":"apple-m-32","label":"Apple M-series 32GB (unified)","vramGb":32,"unified":true,"verdict":"fits","headroomGb":28.52,"totalGb":3.48},{"id":"apple-m-64","label":"Apple M-series 64GB (unified)","vramGb":64,"unified":true,"verdict":"fits","headroomGb":60.52,"totalGb":3.48},{"id":"apple-m-128","label":"Apple M-series 128GB (unified)","vramGb":128,"unified":true,"verdict":"fits","headroomGb":124.52,"totalGb":3.48},{"id":"cpu-ram-32","label":"CPU / 32GB system RAM","vramGb":32,"unified":true,"verdict":"fits","headroomGb":28.52,"totalGb":3.48},{"id":"cpu-ram-64","label":"CPU / 64GB system RAM","vramGb":64,"unified":true,"verdict":"fits","headroomGb":60.52,"totalGb":3.48}],"target":{"id":"h100-80","label":"H100 80GB","vramGb":80,"maxCtx":241664},"whatIf":[{"quant":"q2_k","bytesPerParam":0.33,"weightsGb":23.76,"totalGb":28.35,"verdict":"fits","headroomGb":51.65},{"quant":"q3_k_m","bytesPerParam":0.44,"weightsGb":31.68,"totalGb":36.9,"verdict":"fits","headroomGb":43.1},{"quant":"q4_k_m","bytesPerParam":0.57,"weightsGb":41.04,"totalGb":47.01,"verdict":"fits","headroomGb":32.99},{"quant":"q5_k_m","bytesPerParam":0.69,"weightsGb":49.68,"totalGb":56.34,"verdict":"fits","headroomGb":23.66},{"quant":"q6_k","bytesPerParam":0.82,"weightsGb":59.04,"totalGb":66.45,"verdict":"fits","headroomGb":13.55},{"quant":"q8_0","bytesPerParam":1.06,"weightsGb":76.32,"totalGb":85.11,"verdict":"no","headroomGb":-5.11},{"quant":"f16","bytesPerParam":2,"weightsGb":144,"totalGb":158.2,"verdict":"no","headroomGb":-78.2}],"guidance":["verdict per GPU: fits (<=90% VRAM) / tight (<=100%) / no.","If it does not fit at this precision, check /api/artifacts/{id}/quantizations or the per-quant rows for a smaller mirror.","Estimates: weights + fp16 KV-cache(ctx,batch) + overhead. Real usage varies by runtime."]}