mirror of
https://github.com/hanzo-ml/kserve.git
synced 2026-10-30 20:53:49 +00:00
Dan SunandGitHub
9fb3057831
* Update bert tokenizer example Signed-off-by: Dan Sun <[email protected]> * Generate id for infer request Signed-off-by: Dan Sun <[email protected]> * Fix model name in tests Signed-off-by: Dan Sun <[email protected]> * Fix model_name argument Signed-off-by: Dan Sun <[email protected]> * Fix predictor_protocol Signed-off-by: Dan Sun <[email protected]> * Generate request id if none Signed-off-by: Dan Sun <[email protected]> * Set default None for PredictorConfig Signed-off-by: Dan Sun <[email protected]> * Add comment Signed-off-by: Dan Sun <[email protected]> * Enable doc with default values Signed-off-by: Dan Sun <[email protected]> * Add back add_help false Signed-off-by: Dan Sun <[email protected]> * Update protocol choices Signed-off-by: Dan Sun <[email protected]> * Fix grpc-v2 choice Signed-off-by: Dan Sun <[email protected]> --------- Signed-off-by: Dan Sun <[email protected]>
47 lines
751 B
Protocol Buffer Text Format
47 lines
751 B
Protocol Buffer Text Format
|
|
name: "bert_tf_v2_large_fp16_128_v2"
|
|
platform: "tensorflow_savedmodel"
|
|
max_batch_size: 1
|
|
input [
|
|
{
|
|
name: "unique_ids"
|
|
data_type: TYPE_INT32
|
|
dims: [ 1 ]
|
|
reshape: { shape: [ ] }
|
|
},
|
|
{
|
|
name: "segment_ids"
|
|
data_type: TYPE_INT32
|
|
dims: 128
|
|
},
|
|
{
|
|
name: "input_ids"
|
|
data_type: TYPE_INT32
|
|
dims: 128
|
|
},
|
|
{
|
|
name: "input_mask"
|
|
data_type: TYPE_INT32
|
|
dims: 128
|
|
}
|
|
]
|
|
output [
|
|
{
|
|
name: "end_logits"
|
|
data_type: TYPE_FP32
|
|
dims: 128
|
|
},
|
|
{
|
|
name: "start_logits"
|
|
data_type: TYPE_FP32
|
|
dims: 128
|
|
}
|
|
]
|
|
|
|
instance_group [
|
|
{
|
|
count: 1
|
|
kind: KIND_CPU
|
|
}
|
|
]
|