mirror of
https://github.com/hanzoai/trolley-bench.git
synced 2026-10-30 15:00:49 +00:00
Zachary Landes
8523fe418c
* Labelled weight hashes correctly: 18 files recorded as sha256 held the Hub's Xet hash, so the README's check failed on the right weights; every weights record now carries the true sha256 and, for Xet-stored files, xet_hash, checked against the Hub paths-info API * Made run/weights.py hash every file itself instead of trusting the cache blob name, and added --check to compare a rerun's meta with the committed one by sha256 * Put DECIDER_MODEL and HF_HUB_OFFLINE=1 into the decider-2b command, since its server reads the model from the environment, and had drive.py record the model-selecting variables as server_env * Shipped scenarios/cells-sample.json (324 stratified v1 cells) with its generator, documented --cells, and said plainly that --limit takes the first cells of one scenario and wording * No answer, count, statistic or page file changed; step 1 of the README still rebuilds every derived file byte for byte
51 lines
1.8 KiB
JSON
51 lines
1.8 KiB
JSON
{
|
|
"model_id": "jevk5",
|
|
"code": {
|
|
"repo": "https://github.com/allebee/jevk5",
|
|
"commit": "1e5ae1b533b9eb80c0cbe3fbd010607d0b4e26ae"
|
|
},
|
|
"runtime": "JevK5 v0.3 4B through its documented llama.cpp path (jevk5.gguf.JevK5GGUF, temperature 1.22 / knockout 0.93 from the README table), llama.cpp 9575389 llama-server with Metal, Q8_0 GGUF (quantized: no BF16 GGUF is published and the PyTorch runtime is CUDA-only)",
|
|
"server_command": [
|
|
"./run/jevk5_serve.sh",
|
|
"<scratch>/code/llama.cpp/build/bin/llama-server",
|
|
"<scratch>/hf/hub/models--alibiserikbay--JevK5-GGUF/snapshots/ec67b0bfce5119a8b11a2cdb430bb43e3fa3e82a/jevk5-4b-v0.3-Q8_0.gguf",
|
|
"<scratch>/code/jevk5",
|
|
"8015"
|
|
],
|
|
"startup_s_including_download": 26,
|
|
"probe_answer": {
|
|
"answers": {
|
|
"q": {
|
|
"type": "noul",
|
|
"confidence": 0.9362359061173955,
|
|
"input_tokens": 105,
|
|
"noul": 0.9362359061173955
|
|
}
|
|
},
|
|
"model": "JevK5 v0.3 4B, jevk5-4b-v0.3-Q8_0.gguf"
|
|
},
|
|
"weights": [
|
|
{
|
|
"repo": "alibiserikbay/JevK5-GGUF",
|
|
"revision": "ec67b0bfce5119a8b11a2cdb430bb43e3fa3e82a",
|
|
"bytes": 4482403710,
|
|
"files": {
|
|
"SHA256SUMS": {
|
|
"sha256": "ad21b6c0b7a21fb6cc6980495a91bb2a1c97078beacfc065d4f19e934d6d9422",
|
|
"bytes": 990
|
|
},
|
|
"jevk5-4b-v0.3-Q8_0.gguf": {
|
|
"sha256": "aea433883bc7ed399f2fbd539e53d2eac7caf71a946fe6650995a413979d4a30",
|
|
"xet_hash": "aae363b52d92b29d67ae66fa1990906162f4112373801ebff6326f5197fddc5a",
|
|
"bytes": 4482402720
|
|
}
|
|
}
|
|
}
|
|
],
|
|
"runtime_build": {
|
|
"llama.cpp": "https://github.com/ggml-org/llama.cpp",
|
|
"commit": "9575389609d6f8437de0b205561a4824d217c409",
|
|
"target": "llama-server, Metal"
|
|
},
|
|
"note": "SHA256 of jevk5-4b-v0.3-Q8_0.gguf matches the repo's SHA256SUMS; README table gives temperature 1.22 / knockout 0.93 for this file while the gguf.py docstring says 1.367 for v0.3 4B; the README table value was used"
|
|
} |