baranowskiadam commited on
Commit
1ef7fb5
·
verified ·
1 Parent(s): 41decfc

INT4 DFlash checkpoint weights update

Browse files
Files changed (2) hide show
  1. config.json +13 -1
  2. model.safetensors +1 -1
config.json CHANGED
@@ -21,6 +21,18 @@
21
  "sliding_attention"
22
  ],
23
  "rope_theta": 500000.0,
 
 
 
 
 
 
 
 
 
 
 
 
24
  "gating": "per-head",
25
  "architectures": [
26
  "DFlashLagunaForCausalLM"
@@ -58,4 +70,4 @@
58
  ],
59
  "causal": true
60
  }
61
- }
 
21
  "sliding_attention"
22
  ],
23
  "rope_theta": 500000.0,
24
+ "partial_rotary_factor": 0.5,
25
+ "rope_parameters": {
26
+ "rope_type": "yarn",
27
+ "factor": 32.0,
28
+ "rope_theta": 500000.0,
29
+ "partial_rotary_factor": 0.5,
30
+ "original_max_position_embeddings": 8192,
31
+ "extrapolation_factor": 1.0,
32
+ "attn_factor": 1.0,
33
+ "beta_fast": 32.0,
34
+ "beta_slow": 1.0
35
+ },
36
  "gating": "per-head",
37
  "architectures": [
38
  "DFlashLagunaForCausalLM"
 
70
  ],
71
  "causal": true
72
  }
73
+ }
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:9a8a968a53d78b283d83e5735f9fffec6d3a01f5d1616b7b701b5bc3e76ecd4b
3
  size 2229962896
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0850e39b5c079a9f1a9bafed729a4545b088a91876541d010d871f6d6d8bf909
3
  size 2229962896