MARIO-Math-Reasoning Super_MARIO issues

MARIO-Math-Reasoning / Super_MARIO

MIT License

172 stars 13 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

Issue about requirements file

#18 LePanda026 closed 1 week ago
3
type of template for training

#17 vgaraujov closed 1 month ago
7
About Training data generation.

#16 George-Chia closed 2 weeks ago
2
Possible problems about the training dataset

#15 FlyingDutchman26 closed 2 months ago
2
Why are the batch size and number of epochs much larger than common SFT settings?

#14 tongyx361 closed 2 months ago
3
Is the model initialized from pre-trained model or model from the last iteration round for each round?

#13 tongyx361 closed 2 months ago
2
Why not directly generate the value, but instead add a value head? Could you explain the reasoning behind this decision?

#12 yanzhenqiang closed 3 months ago
1
value estimation twice?

#11 platoonpluto closed 3 months ago
5
AttributeError: 'RequestOutput' object has no attribute 'value_estimate'

#10 yanzhenqiang closed 3 months ago
1
MCTS training data generation in round1

#9 platoonpluto closed 3 months ago
1
training code

#8 jordane95 closed 3 months ago
2
How to set B1 in Step level Beam Search

#7 xiaolizh1 closed 3 months ago
3
How to initialize first generation child nodes?

#6 Jeff123z closed 3 months ago
1
Update solver_demo.py

#5 eltociear closed 3 months ago
0
数学推理本身是个非对称二元博弈问题

#4 hxypqr closed 3 months ago
3
About the code

#3 liushz closed 3 months ago
3
AlphaMath listed as AlaphaMath in Huggingface

#2 1of13 closed 3 months ago
1
Concern on (first few rounds) sampling efficacy

#1 billxbf closed 4 months ago
3