Sage
predict_performance
$0.05 USDC: predict configuration throughput, memory, and power from bounded StackBench evidence with confidence, missing metrics, and caveats.
Remote computesage/stackbench
Remote (network-hosted)
Other tools also called predict_performance?
See providers with this name
Input Schema
{
"type": "object",
"properties": {
"model": {
"type": "string",
"maxLength": 240,
"minLength": 1
},
"engine": {
"anyOf": [
{
"type": "string",
"maxLength": 240,
"minLength": 1
},
{
"type": "null"
}
]
},
"batch_size": {
"anyOf": [
{
"type": "integer",
"maximum": 9007199254740991,
"exclusiveMinimum": 0
},
{
"type": "null"
}
]
},
"concurrency": {
"type": "integer",
"default": 1,
"maximum": 9007199254740991,
"exclusiveMinimum": 0
},
"ubatch_size": {
"anyOf": [
{
"type": "integer",
"maximum": 9007199254740991,
"exclusiveMinimum": 0
},
{
"type": "null"
}
]
},
"max_evidence": {
"type": "integer",
"default": 12,
"maximum": 50,
"minimum": 1
},
"quantization": {
"anyOf": [
{
"type": "string",
"maxLength": 240,
"minLength": 1
},
{
"type": "null"
}
]
},
"prompt_tokens": {
"type": "integer",
"maximum": 9007199254740991,
"exclusiveMinimum": 0
},
"engine_version": {
"anyOf": [
{
"type": "string",
"maxLength": 240,
"minLength": 1
},
{
"type": "null"
}
]
},
"flash_attention": {
"anyOf": [
{
"type": "boolean"
},
{
"type": "null"
}
]
},
"hardware_system": {
"type": "string",
"maxLength": 240,
"minLength": 1
},
"kv_cache_type_k": {
"anyOf": [
{
"type": "string",
"maxLength": 240,
"minLength": 1
},
{
"type": "null"
}
]
},
"kv_cache_type_v": {
"anyOf": [
{
"type": "string",
"maxLength": 240,
"minLength": 1
},
{
"type": "null"
}
]
},
"requested_metrics": {
"type": "array",
"items": {
"enum": [
"prefill_tok_s",
"decode_tok_s",
"peak_vram_bytes",
"avg_power_watts"
],
"type": "string"
},
"default": [
"prefill_tok_s",
"decode_tok_s",
"peak_vram_bytes",
"avg_power_watts"
],
"maxItems": 4,
"minItems": 1
},
"decode_context_depth": {
"anyOf": [
{
"type": "integer",
"maximum": 9007199254740991,
"minimum": 0
},
{
"type": "null"
}
]
}
}
}