Add gemma support #69
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
qihqi
force-pushed
the
hanq_add_model
branch
2 times, most recently
from
b77f88c to
eb739eb
Compare
shard config for llamA gemma 3 gemma4 formatter
FanhaiLu1
approved these changes
May 4, 2024
| # "replicated" to signify "replicated". | ||
| # Integer signify axis to shard: 0 <= shard axis < rank | ||
|
|
||
| freqs_cis : null # torch.complex64 (16384, 128) |
Collaborator
There was a problem hiding this comment.
Can we keep consistency on replicated sharding? If either null or -1 is fine, shall we just keep -1 in our code base (use null in gemma, but -1 in llama)?
| ) | ||
| return caches | ||
|
|
||
| def sharding_by_name(self, name): |
Collaborator
There was a problem hiding this comment.
Great, it's more clear than previous hardcode one.
| """Attention module.""" | ||
|
|
||
| def __init__(self, args, env): | ||
| def __init__(self, n_heads, n_kv_heads, head_dim, hidden_size, device, env): |
Collaborator
There was a problem hiding this comment.
Great, it's nice to see the layers is decoupled with args.
| print(f"---------> {jax.devices()}") | ||
|
|
||
| env, model_arg = helpers.make_env_tiny(bf16_enable=False) | ||
| torch.set_default_dtype(torch.float32) |
Collaborator
There was a problem hiding this comment.
We can remove this line. helpers.make_env_tiny has code to do set_default_dtype:
torch_dtype = torch.bfloat16 if bf16_enable else torch.float32
torch.set_default_dtype(torch_dtype)
Collaborator
There was a problem hiding this comment.
same as other torch.set_default_dtype in this files.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Changes include:
Run interactive and run server both works with random weights.