Hugging Face's logo

runanywhere
/
nemotron_nano_8b_HNPU

Text Generation
hnpu
hexagon
npu
llm
v81
Model card Files
xet
 Community
1
sanmonga22 commited on
Commit
fd6169c
·
verified ·
1 Parent(s): b6a26fb

Restore v75 QHexRT HNPU artifacts

Browse files

Restore only the deleted v75/ runtime artifacts needed for Hexagon v75 support. No non-v75 cleanup artifacts are restored. v75 manifests are restored with non-runtime _comment metadata stripped.

Files changed (12) hide show
  1. v75/nemotron-nano-8b.json +66 -0
  2. v75/nemotronnano8b_decode_p0_w8.bin +3 -0
  3. v75/nemotronnano8b_decode_p1_w8.bin +3 -0
  4. v75/nemotronnano8b_decode_p2_w8.bin +3 -0
  5. v75/nemotronnano8b_decode_p3_w8.bin +3 -0
  6. v75/nemotronnano8b_decode_p4_w8.bin +3 -0
  7. v75/nemotronnano8b_decode_p5_w8.bin +3 -0
  8. v75/nemotronnano8b_decode_p6_w8.bin +3 -0
  9. v75/nemotronnano8b_decode_p7_w8.bin +3 -0
  10. v75/nemotronnano8b_embed_f16.bin +3 -0
  11. v75/nemotronnano8b_lmhead_w8.bin +3 -0
  12. v75/tokenizer.json +3 -0
v75/nemotron-nano-8b.json ADDED
@@ -0,0 +1,66 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema_version": 1,
3
+ "model": {
4
+ "name": "nemotron-nano-8b",
5
+ "family": "llm",
6
+ "dsp_arch": "v75",
7
+ "tokenizer_pre": "llama3"
8
+ },
9
+ "params": {
10
+ "hidden": 4096,
11
+ "vocab": 128256,
12
+ "n_layers": 32,
13
+ "max_ctx": 512,
14
+ "kv_dim": 1024,
15
+ "head_dim": 128,
16
+ "rope_theta": 500000.0,
17
+ "eos_token_id": 128001
18
+ },
19
+ "artifacts": {
20
+ "contexts": {
21
+ "decode_p0": {
22
+ "bin": "nemotronnano8b_decode_p0_w8.bin"
23
+ },
24
+ "decode_p1": {
25
+ "bin": "nemotronnano8b_decode_p1_w8.bin"
26
+ },
27
+ "decode_p2": {
28
+ "bin": "nemotronnano8b_decode_p2_w8.bin"
29
+ },
30
+ "decode_p3": {
31
+ "bin": "nemotronnano8b_decode_p3_w8.bin"
32
+ },
33
+ "decode_p4": {
34
+ "bin": "nemotronnano8b_decode_p4_w8.bin"
35
+ },
36
+ "decode_p5": {
37
+ "bin": "nemotronnano8b_decode_p5_w8.bin"
38
+ },
39
+ "decode_p6": {
40
+ "bin": "nemotronnano8b_decode_p6_w8.bin"
41
+ },
42
+ "decode_p7": {
43
+ "bin": "nemotronnano8b_decode_p7_w8.bin"
44
+ },
45
+ "lmhead": {
46
+ "bin": "nemotronnano8b_lmhead_w8.bin"
47
+ }
48
+ },
49
+ "embed": "nemotronnano8b_embed_f16.bin",
50
+ "tokenizer": "tokenizer.json"
51
+ },
52
+ "plan": {
53
+ "steps": [
54
+ {
55
+ "host": "llama_generate_sharded",
56
+ "params": {
57
+ "decode_parts": "nemotronnano8b_decode_p0_w8,nemotronnano8b_decode_p1_w8,nemotronnano8b_decode_p2_w8,nemotronnano8b_decode_p3_w8,nemotronnano8b_decode_p4_w8,nemotronnano8b_decode_p5_w8,nemotronnano8b_decode_p6_w8,nemotronnano8b_decode_p7_w8",
58
+ "lmhead": "nemotronnano8b_lmhead_w8",
59
+ "max_new": 64,
60
+ "neg": -50000,
61
+ "rope_factor": 8.0
62
+ }
63
+ }
64
+ ]
65
+ }
66
+ }
v75/nemotronnano8b_decode_p0_w8.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7b66e1eef4bc1903942bf09cb4823eddd9df57d20ea25d59244c4fa8d70eed76
3
+ size 876392448
v75/nemotronnano8b_decode_p1_w8.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0b0a9aab17bbd5c27a433dc267011368279d6d14e1949c4c44af2d18e458e0df
3
+ size 876392448
v75/nemotronnano8b_decode_p2_w8.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ca9a95042ae0e5f259174faca2cf14bd47b8169bd96c1e1d3771659773c6016b
3
+ size 876392448
v75/nemotronnano8b_decode_p3_w8.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a582d7bd4dbe8907ebe3e763329d8d3da7c5b58ea9ea1cc42d3262a4d871684a
3
+ size 876392448
v75/nemotronnano8b_decode_p4_w8.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:eb1463fcbdb1cedd3c7b7d181fc20e3c7d28541d00a7ff8d6d90d5b4ec23d3a0
3
+ size 876392448
v75/nemotronnano8b_decode_p5_w8.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f5d1a6ea2a9fc086992b2e87f672c1d6bd8e3ae4cf5ec8078efcdeb69810bb5f
3
+ size 876392448
v75/nemotronnano8b_decode_p6_w8.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:543a315646fd0803230111d06f8de370bb5a6d7ab3efaf55b76fd4fb647852b8
3
+ size 876392448
v75/nemotronnano8b_decode_p7_w8.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ffbfd021e63a2bdd0cae5e904a47b19125c4a63f0fa3299d91ef7b13dbc2ac24
3
+ size 876400640
v75/nemotronnano8b_embed_f16.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8443236c0685bf39fb998183ecd98d14e296cb6fee1e1c7841ec89d2c15fcfac
3
+ size 1050673152
v75/nemotronnano8b_lmhead_w8.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:844a917ae15d026d9b418c04e9726a0619f90208aca7a5987d2b80841353cb8a
3
+ size 527147008
v75/tokenizer.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6b9e4e7fb171f92fd137b777cc2714bf87d11576700a1dcd7a399e7bbe39537b
3
+ size 17209920