examples of new GPT2 and JIT change (#2261)

* var_vals are global

* working with global ish

* better

* fix export model

* fix tests

* better kv cache

* does it run?

* use where for kvmask

* fix excessive var_vals

* fix import

* how does multigpu use this?

* llama kinda work

* faster and simpler

* cleanup

* fix conversation mode

* test cleanups

* fix one more test

* test cleanup

---------

Co-authored-by: George Hotz <geohot@gmail.com>
This commit is contained in:
chenyu
2023-11-10 15:07:02 -05:00
committed by GitHub
parent b6aaf12df7
commit a753c8e071
15 changed files with 189 additions and 249 deletions

View File

@@ -300,10 +300,9 @@ cache_saved = CacheCollector.finish() # disable the cache
# there's one ASTRunner in the cache
assert len(cache_saved) == 1
prg, bufs, _ = cache_saved[0]
# print the C Program :)
print(prg.prg)
print(cache_saved[0].prg.prg)
# after some formatting (the compiler doesn't care)
# NOTE: the 2 and 3 are constant folded