Validation step3 #21
Open
Validation step3 #21
Conversation
This file contains 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
…ectly (like dtype != float, requires grad), make flash attention dep optional (still needs to be checked)
AleHD
requested changes
Nov 21, 2024
There was a problem hiding this comment.
Looks good to me. Was able to do a minimal test after the dp>1 fix. Have you tried saving and loading checkpoints to make sure valid dataloader samples are reproducible?
Also, I'd vote to add a way to held out a portion of the train set and use it as validation, instead of needing to add an additional dataset necessarily. Something like the --split argument in megatron https://github.com/NVIDIA/Megatron-LM/blob/81fee9b0047fb3ac6001b5e71e4df89fc01b2a1c/megatron/training/arguments.py#L1760-L1764. How hard would it be to implement this?
Comment on lines
+1
to
+91
| import torch | ||
| import torchtune | ||
| import flash_attn | ||
| import flash_attn.layers.rotary | ||
|
|
||
|
|
||
| class RotaryEmbeddingKyleLikeFA(torch.nn.Module): | ||
| """ | ||
| Has the same function signature as FA, for interleaved=True and separate q, kv. | ||
| seqlen_offset = 0 | ||
| Does not operate inplace, but that's fine for how it's used in Nanotron. | ||
| """ | ||
| def __init__(self, dim: int, base: float): | ||
| super().__init__() | ||
| self.dim = dim | ||
| self.base = float(base) | ||
|
|
||
| self.max_seq_len = None | ||
| self.rpe = None | ||
|
|
||
| def forward(self, q, kv): | ||
| bs, q_len, n_heads, _ = q.shape | ||
| assert self.dim == _ | ||
|
|
||
| assert (bs, q_len, 2, n_heads, self.dim) == kv.shape | ||
|
|
||
| if (self.rpe is None) or (self.max_seq_len != q_len): | ||
| self.max_seq_len = q_len | ||
| self.rpe = torchtune.modules.RotaryPositionalEmbeddings(dim=self.dim, | ||
| max_seq_len=self.max_seq_len, | ||
| base=self.base).to(device) | ||
| q_out = self.rpe(q) | ||
| kv_out = torch.stack((self.rpe(kv[:, :, 0]), kv[:, :, 1]), 2) | ||
| return q_out, kv_out | ||
|
|
||
|
|
||
|
|
||
| if __name__ == "__main__": | ||
| device = torch.device(0) | ||
| theta = 10000 | ||
|
|
||
| batch_size = 3 | ||
| dim_qk = 4 | ||
| q_len = 256 | ||
| kv_len = 256 | ||
| n_heads = 4 | ||
|
|
||
| max_seq_len = max(q_len, kv_len) | ||
|
|
||
| print(max_seq_len) | ||
|
|
||
|
|
||
| query_states = torch.rand(batch_size, q_len, n_heads, dim_qk, device=device) | ||
| key_value_states = torch.rand(batch_size, kv_len, 2, n_heads, dim_qk, device=device).contiguous() | ||
|
|
||
|
|
||
| interleaved = True | ||
| # interleaved = False | ||
| re1 = flash_attn.layers.rotary.RotaryEmbedding(dim=dim_qk, interleaved=interleaved, base=theta).to(device) | ||
| re2 = torchtune.modules.RotaryPositionalEmbeddings(dim=dim_qk, max_seq_len=max_seq_len, base=theta).to(device) | ||
| re3 = RotaryEmbeddingKyleLikeFA(dim=dim_qk, base=theta).to(device) | ||
|
|
||
|
|
||
|
|
||
| print(key_value_states[:, :, 0].shape) | ||
|
|
||
| out2 = re2(query_states) | ||
| out3 = re2(key_value_states[:, :, 0]) | ||
| # out4 = re2(key_value_states[:, :, 1]) | ||
|
|
||
| out_eq = re3(query_states, kv=key_value_states) | ||
|
|
||
| # torch.testing.assert_close(out2, query_states) | ||
| out1 = re1(query_states, kv=key_value_states) | ||
|
|
||
| torch.testing.assert_close(out_eq[0], out1[0]) | ||
| torch.testing.assert_close(out_eq[1], out1[1]) | ||
|
|
||
|
|
||
| # Do this second, since the computation is inplace | ||
| torch.testing.assert_close(out1[0], query_states) | ||
|
|
||
| test = torch.stack((out3, key_value_states[:, :, 1]), 2) | ||
| torch.testing.assert_close(out1[1], test) | ||
| # torch.testing.assert_close(out1[1][:, :, 0], out3) | ||
|
|
||
|
|
||
| torch.testing.assert_close(out1[0], out2) | ||
|
|
||
| print("done") | ||
|
|
There was a problem hiding this comment.
I'd say this one should either go somewhere to tests or be removed.
Co-authored-by: AleHC <[email protected]>
Co-authored-by: AleHC <[email protected]>
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.
No description provided.