Hugging Face's logo

ecker
/
vall-e

Model card Files Community
ecker commited on
Commit
343402b
•
1 Parent(s): 8642a4f

Upload 2 files

Browse files

Updated reference AR+NAR model to a "better" trained one (need to do actual regression tests, but it sounds really nice), backed up current one as `ar+nar-old-llama-8`

Files changed (3) hide show
  1. .gitattributes +1 -0
  2. models/ckpt/ar+nar-llama-8/fp32.sft +2 -2
  3. models/ckpt/ar+nar-old-llama-8/fp32.sft +3 -0
.gitattributes CHANGED
@@ -19,3 +19,4 @@ loras/ckpt/lora-portal-glados-r128-a128/lora.sft filter=lfs diff=lfs merge=lfs -
19
  loras/ckpt/lora-samandmax-sam-r128-a128/lora.sft filter=lfs diff=lfs merge=lfs -text
20
  models/ckpt/ar+nar-layerskip-llama-8/fp32.sft filter=lfs diff=lfs merge=lfs -text
21
  models/ckpt/nar-len-llama-8/fp32.sft filter=lfs diff=lfs merge=lfs -text
 
 
19
  loras/ckpt/lora-samandmax-sam-r128-a128/lora.sft filter=lfs diff=lfs merge=lfs -text
20
  models/ckpt/ar+nar-layerskip-llama-8/fp32.sft filter=lfs diff=lfs merge=lfs -text
21
  models/ckpt/nar-len-llama-8/fp32.sft filter=lfs diff=lfs merge=lfs -text
22
+ models/ckpt/ar+nar-old-llama-8/fp32.sft filter=lfs diff=lfs merge=lfs -text
models/ckpt/ar+nar-llama-8/fp32.sft CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:0920e7eef0884631f00513b700924538db0853d662530e4bcf7ac1d8666430b6
3
- size 456274402
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ad5fc4b3f6ef8c847b3fd0edd93154b92f51b54b1645fcf1a78a9d35dbb3c38f
3
+ size 454198146
models/ckpt/ar+nar-old-llama-8/fp32.sft ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0920e7eef0884631f00513b700924538db0853d662530e4bcf7ac1d8666430b6
3
+ size 456274402