train parameters are incorrect in README file

[bluedevils23]

Usage:
trainer configuration [OPTION...]

-h, --help Print help
-l, --log_level arg log level (default: info)
-f, --log_file file?
-c, --cfr arg MUST: file to the cfr file
-b, --builder arg MUST: file to the abstract game def file
-s, --sgs arg MUST: file to the sgs file
-e, --exp_tag arg tag to persist the lab file (default: x)
-p, --load_checkpoint arg loading strategy checkpoint
-t, --blueprint arg loading blueprint for depth limit solving
-m, --match_state arg root state to start training from

It is totally different with README file.

[dpmaloney]

Tons is missing from the codebase/is out of date tbh, I don't think you can even use a hierarchical abstraction builder because theres no "hand_histo" files nor is there a way to generate them.

If you wanna collaborate to clean it up I'd be happy to work on a fork with you.

[bluedevils23]

Need to generate buckets first
./build_buckets --mode=hierarchical --private_cards=2 --public_cards=4 --clusters=200 --threads=24 --loglevel=info
2021-12-22 12:56:17|11180|critical|missing /bulldog/data/abstraction/buckets_200_2_4.bin

Needs to figure out how to generate bin file first.

[kmurf1999]

@dpmaloney @bluedevils23 I've done hand histogram generation and clustering in rust (https://github.com/kmurf1999/CSE485PokerSolver/blob/master/poker_solver/information_abstraction/src/bin/gen_hist.rs)
If any of you would like to collaborate on this, let me know.

[dpmaloney]

More than happy, there’s some new papers that I’ve been implementing as well that might be interesting to you guys.

[parmercapital]

@dpmaloney did you ever get around to a working fork? Trying to spin this up and running into a lot of issues.

[calb-harb]

Also interested in knowing if anyone else has continued this work

[cubesnyc]

@kmurf1999 are you still interested in collaborating on this? whats the best way to get in touch?