初始化项目,由ModelHub XC社区提供模型
Model: TaH-plus/qwen3-1.7b-tah-duo-maxiter4-amteam121k-sft Source: Original Platform
This commit is contained in:
File diff suppressed because one or more lines are too long
@@ -0,0 +1,37 @@
|
||||
_logger_level: WARNING
|
||||
_random_seed: 42
|
||||
batch_size: 8
|
||||
dtype: bfloat16
|
||||
eval_iter_decider: iter_decider
|
||||
eval_iter_decider_kwargs: {}
|
||||
iter_decider: Qwen3MLPIterDecider
|
||||
iter_decider_kwargs:
|
||||
hidden_size: 2048
|
||||
intermediate_size: 2048
|
||||
max_iter: 2
|
||||
threshold: 0.5
|
||||
topk: 2048
|
||||
max_iter: 2
|
||||
max_new_tokens: 16384
|
||||
mem_fraction_static: 0.8
|
||||
min_p: 0.0
|
||||
online_backend: mini_sglang_online
|
||||
online_base_urls:
|
||||
- http://127.0.0.1:30080
|
||||
- http://127.0.0.1:30081
|
||||
- http://127.0.0.1:30082
|
||||
- http://127.0.0.1:30083
|
||||
- http://127.0.0.1:30084
|
||||
- http://127.0.0.1:30085
|
||||
- http://127.0.0.1:30086
|
||||
- http://127.0.0.1:30087
|
||||
per_server_concurrency: 16
|
||||
prompt_iter_count: 1
|
||||
repeat_size: 32
|
||||
temperature: 0.6
|
||||
top_k: 20
|
||||
top_p: 0.95
|
||||
total_concurrency: 128
|
||||
tracker_kwargs:
|
||||
top_k: 5
|
||||
use_tracker: false
|
||||
@@ -0,0 +1,33 @@
|
||||
problem_id,accuracy,correct_count,total_samples,avg_output_length
|
||||
aime25_0,0.844,27,32,16384.00
|
||||
aime25_1,0.062,2,32,16383.78
|
||||
aime25_2,0.688,22,32,16384.00
|
||||
aime25_3,0.344,11,32,16384.00
|
||||
aime25_4,0.000,0,32,16383.97
|
||||
aime25_5,0.812,26,32,16383.97
|
||||
aime25_6,0.000,0,32,16383.97
|
||||
aime25_7,0.031,1,32,16384.00
|
||||
aime25_8,0.031,1,32,16383.94
|
||||
aime25_9,0.000,0,32,16384.00
|
||||
aime25_10,0.000,0,32,16383.69
|
||||
aime25_11,0.000,0,32,16384.00
|
||||
aime25_12,0.000,0,32,16384.00
|
||||
aime25_13,0.062,2,32,16383.97
|
||||
aime25_14,0.000,0,32,16384.00
|
||||
aime25_15,0.844,27,32,16384.00
|
||||
aime25_16,0.250,8,32,16384.00
|
||||
aime25_18,0.281,9,32,16383.88
|
||||
aime25_17,0.000,0,32,16383.75
|
||||
aime25_19,0.000,0,32,16384.00
|
||||
aime25_20,0.031,1,32,16384.00
|
||||
aime25_21,0.000,0,32,16384.00
|
||||
aime25_22,0.000,0,32,16384.00
|
||||
aime25_23,0.031,1,32,16384.00
|
||||
aime25_24,0.000,0,32,16384.00
|
||||
aime25_25,0.000,0,32,16384.00
|
||||
aime25_26,0.000,0,32,16384.00
|
||||
aime25_27,0.000,0,32,16383.91
|
||||
aime25_28,0.031,1,32,16384.00
|
||||
aime25_29,0.000,0,32,16384.34
|
||||
|
||||
Total Accuracy,0.145,139,960,16383.97
|
||||
|
@@ -0,0 +1,7 @@
|
||||
iter_count,token_count,percentage
|
||||
1,10226087,65.0157101949056
|
||||
2,1528303,9.716688791910807
|
||||
3,732614,4.657834370930989
|
||||
4,3241636,20.609766642252605
|
||||
|
||||
Total Tokens,15728640,100.0
|
||||
|
@@ -0,0 +1,16 @@
|
||||
{
|
||||
"pass@32": 0.4667,
|
||||
"cons@32": 0.2333,
|
||||
"avg@32": 0.1448,
|
||||
"avg_output_length": 16383.97,
|
||||
"num_problems": 30,
|
||||
"num_samples": 960,
|
||||
"repeat_size": 32,
|
||||
"avg_request_time": 325.8334,
|
||||
"end_2_end_time": 2611.8533,
|
||||
"token_throughtput": 6022.0123,
|
||||
"request_throught": 0.3676,
|
||||
"avg_iter_count": 1.8086,
|
||||
"total_concurrency": 128,
|
||||
"failed_requests": 0
|
||||
}
|
||||
@@ -0,0 +1,3 @@
|
||||
version https://git-lfs.github.com/spec/v1
|
||||
oid sha256:5780e1a7b0958b804ac6c1d7c4e04b743248d31033b8c98eeccc5e2cc8d6d0e7
|
||||
size 48236268
|
||||
Reference in New Issue
Block a user