qwenti-inference / results.json
OpceanAI's picture
Model card, paper, figures, logs and scripts
2049fc8 verified
Raw History Blame Contribute Delete
3.61 kB
{
"sections": {
"hardware": {
"gpu": "NVIDIA RTX PRO 6000 Blackwell Server Edition",
"vram_gb": 95.6,
"driver": "580.82.07",
"torch": "2.14.1+cu130",
"cuda": "13.0",
"capability": "12.0",
"cpus": 48,
"ram_gb": 176.9,
"python": "3.13.16"
},
"setup": {
"export": "bash: line 1: cd: /content/tl: No such file or directory",
"rc": [
0,
1,
1,
1
]
},
"model": {
"error": "CalledProcessError: Command '['cp', '/root/.cache/huggingface/hub/models--Maykeye--TinyLLama-v0/snapshots/298338802ab94432b917bcce11382aa151aee50f/model.safetensors', '/content/tl/model.safetensors']' returned non-zero exit status 1.",
"tb": "Traceback (most recent call last):\n File \"/tmp/ipykernel_1496/361909880.py\", line 23, in deco\n R[\"sections\"][name] = fn()\n ~~^^\n File \"/tmp/ipykernel_1496/361909880.py\", line 73, in model\n subprocess.run([\"cp\", f\"{p}/{f}\", f\"/content/tl/{f}\"], check=True)\n ~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\n File \"/usr/lib/python3.13/subprocess.py\", line 578, in run\n raise CalledProcessError(retcode, process.args,\n output=stdout, stderr=stderr)\nsubprocess.CalledProcessError: Command '['cp', '/root/.cache/huggingface/hub/models--Maykeye--TinyLLama-v0/snapshots/298338802ab94432b917bcce11382aa151aee50f/model.safetensors', '/content/tl/model.safetensors']' returned non-zero exit status 1.\n"
},
"trace": {
"error": "ModuleNotFoundError: No module named 'chipformer'",
"tb": "Traceback (most recent call last):\n File \"/tmp/ipykernel_1496/361909880.py\", line 23, in deco\n R[\"sections\"][name] = fn()\n ~~^^\n File \"/tmp/ipykernel_1496/361909880.py\", line 89, in trace\n from chipformer.rv import elf, ref\nModuleNotFoundError: No module named 'chipformer'\n"
},
"executor": {
"error": "ModuleNotFoundError: No module named 'chipformer'",
"tb": "Traceback (most recent call last):\n File \"/tmp/ipykernel_1496/361909880.py\", line 23, in deco\n R[\"sections\"][name] = fn()\n ~~^^\n File \"/tmp/ipykernel_1496/361909880.py\", line 112, in executor\n from chipformer.rv import core, elf, fastexec, ref\nModuleNotFoundError: No module named 'chipformer'\n"
},
"merge": {
"rc": 0,
"size_mb": 27408,
"circuit_layers": null,
"tail": "use `flash-linear-attention` is not installed. This is correct but much slower; install `flash-linear-attention` for the optimized kernel.\n{\"hidden\": 4096, \"chip_layers\": 36, \"slots\": 1457, \"circuit\": {\"layers\": 36, \"dims\": 3259, \"heads\": 457, \"ffn_units\": 46685, \"max_units\": 8648, \"max_heads\": 16}}"
},
"degradation": {
"per_text": [
{
"kw": 0.000528,
"top1": 1.0,
"positions": 18
},
{
"kw": 0.000782,
"top1": 1.0,
"positions": 21
},
{
"kw": 0.000657,
"top1": 1.0,
"positions": 14
},
{
"kw": 0.000616,
"top1": 1.0,
"positions": 18
},
{
"kw": 0.000701,
"top1": 1.0,
"positions": 13
},
{
"kw": 0.000621,
"top1": 1.0,
"positions": 13
},
{
"kw": 0.000808,
"top1": 1.0,
"positions": 21
},
{
"kw": 0.000566,
"top1": 1.0,
"positions": 17
}
],
"kl_mean": 0.00066,
"positions": 135,
"sample_continuation": "!\nA. London\nB. Paris\nC. Berlin\nD. Madrid\n\n<think>\n",
"decode_tok_s": 23.86,
"peak_vram_gb": 34.7,
"base_layers": 32,
"merged_layers": 68
}
}
}