Mirrors/text-generation-webui
1
0
Fork 0
mirror of https://github.com/oobabooga/text-generation-webui.git synced 2024-12-25 05:48:55 +01:00
Code Issues Packages Projects Releases Wiki Activity

Revert "Fix exllama_hf gibbersh above 2048 context, and works >5000 context. (#2913)"

Browse Source 
This reverts commit 37a16d23a7.
This commit is contained in:
oobabooga 2023-06-28 18:10:34 -03:00
parent 7b048dcf67
commit 20740ab16e
1 changed files with 1 additions and 9 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

10
modules/exllama_hf.py
View File

@ -54,15 +54,7 @@ class ExllamaHF(PreTrainedModel):
cache = kwargs['past_key_values'] if 'past_key_values' in kwargs else None
if cache is None:
cache = ExLlamaCache(self.ex_model)
nseq = seq[:-1]
for seqs in [nseq[i : i + 2048] for i in range(0, len(nseq), 2048)]:
self.ex_model.forward(
torch.tensor([seqs], dtype=torch.long),
cache,
preprocess_only=True,
lora=self.lora,
)
self.ex_model.forward(torch.tensor([seq[:-1]], dtype=torch.long), cache, preprocess_only=True, lora=self.lora)
logits = self.ex_model.forward(torch.tensor([seq[-1:]], dtype=torch.long), cache, lora=self.lora).to(kwargs['input_ids'].device)