* Slice to unbind symbolic
* use vmax for now
* assert shape in reshape is valid
* update test_symbolic_ops to use shrink instead of reshape
* remove infer_with_bound_values for npw
* symbolic output doesnt have symbolic strides
* symbolic jit tests use shrink to unregister symbolic
* update test
* update more tests
* wrap vmax in int()
* only create a new st if the store is not an assigne
* unwrap st
* comments
* Modify tests and start work towards removing symbolic reshape
* Refactor symbolic reshape
* fix small error
* much cleaner + fix more tests
* Can remove this now
* Update test_symbolic_ops and test_tiny
* Couple more tests
* Unused import
* More tests and add EXPAND to Tensor.empty
* Fix test beam search
* all int
* Fix rangeify by adding shrink
* Remove OOB check and so fix test_symbolic_jit
* test_symbolic_jit doesn't need OOB Context anymore either
* Should remove that test now
* Cleanups part 1
* fix linters
* Final cleanups
* Don't reassign inside for loop
---------
Co-authored-by: chenyu <chenyu@fastmail.com>
* support symbolic reshape with non-contiguous
pre-requisite for symbolic arange (make symbolic ones that can be folded).
* test cases
* typo
* shorter
currently not supporting const fold symbolic shape. I think it's possible with a refactor to Tensor.from_node.
also added some failed required tests for symbolic arange.
* add symbolic mean test cases in test_symbolic_ops and test_symbolic_jit
2d symbolic mean in jit does not quite work, order of the variable inputs are not deterministic?
* skip
* lazy rewrite, try 2
* min fix tests
* pass contig test
* put broken pads back
* move that to realize
* no contig child fixes array packing
* so wrong
* now that's correct
* base children
* fix bind issues
* disable to_image_idx
* fix tests
* that failure shouldn't break other tests
* more fixes
* fix torch
* skip failing tests in CI
* 1e-7
* half is broken
* 1e-6 margin of error
* 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>
* no JIT call in TransformerBlock
* idea
* move 2 reshapes to jitted function
shrink inside jitted too, 6.3ms
remove back reshapes, 5.5ms
isinstance -> __class__ 4.99ms
* think
revert ops_gpu.py
revert symbolic.py too
PYOPENCL_COMPILER_OUTPUT=1
* cleanup
* fix cache shape for conversational model
only reshape if start_pos > 0
* small cleanup
* include var_vals.keys() to st.key
* add comments
* llama small update
* everything jitted again, similar structure to gpt2
* fix typing
* add TODO for in place update cache
* Symbolic Shape JIT
update tests
2 variables symbolic ops, adding more tests
test passing
cleanup
* more test cases
* single flag
* review update
* jit attention one piece
* realize
* symbolic_jit test for cuda
* old artifact
* works with cuda gpu but failed ci
* CUDACPU
