Upload folder using huggingface_hub

.gitattributes

@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+v81/tokenizer.json filter=lfs diff=lfs merge=lfs -text

v81/README.md

@@ -0,0 +1,10 @@
+---
+license: other
+base_model: nvidia/Llama-3.1-NemoGuard-8B-Topic-Control
+tags: [qualcomm, hexagon, npu, qnn, qhexrt, v81, llama, guardrail, topic-control]
+---
+# Llama-3.1-NemoGuard-8B-Topic-Control — Hexagon NPU (v81)
+QHexRT-runnable, arch-pinned QNN bundle of NVIDIA's NemoGuard topic-control guardrail for **Hexagon v81**
+(SM8850, soc_model 87). LoRA on Llama-3.1-8B-Instruct, merged onto an ungated base, converted via QHexRT's
+**dense decode-split** (8 chained `decode_p*` parts + `llama_generate_sharded`). **Device-validated W8
+greedy-EXACT vs HF gold @ ~3.6 tok/s.** See `runanywhere/nemotron_nano_8b_HNPU` for the run recipe.

v81/nemoguard-topic-8b.json

@@ -0,0 +1,67 @@
+{
+  "schema_version": 1,
+  "_comment": "Llama-3.1-Nemotron-Nano-8B-v1 (NVIDIA reasoning fine-tune of Llama-3.1-8B) on Hexagon v81 (SM8850 / soc_model 87). Standard dense Llama-3.1-8B (32 q heads -> materialized MHA + MAXCTX 512; llama3-scaled RoPE theta 500000 factor 8; untied lm-head). The ~7GB W8 decode exceeds the v81 cDSP per-context ceiling, so the 32 layers are split into 8 CHAINED decode parts (decode_parts=8, ~0.85GB W8 each) driven by the llama_generate_sharded host-op (decode-over-prompt seed, no batched prefill). Export gate: decode 8/8 match gold (forge). W8 device gate is greedy_tol.",
+  "model": {
+    "name": "nemoguard-topic-8b",
+    "family": "llm",
+    "dsp_arch": "v81",
+    "tokenizer_pre": "llama3"
+  },
+  "params": {
+    "hidden": 4096,
+    "vocab": 128256,
+    "n_layers": 32,
+    "max_ctx": 512,
+    "kv_dim": 1024,
+    "head_dim": 128,
+    "rope_theta": 500000.0,
+    "eos_token_id": 128001
+  },
+  "artifacts": {
+    "contexts": {
+      "decode_p0": {
+        "bin": "nemoguardtopic8b_decode_p0_w8.bin"
+      },
+      "decode_p1": {
+        "bin": "nemoguardtopic8b_decode_p1_w8.bin"
+      },
+      "decode_p2": {
+        "bin": "nemoguardtopic8b_decode_p2_w8.bin"
+      },
+      "decode_p3": {
+        "bin": "nemoguardtopic8b_decode_p3_w8.bin"
+      },
+      "decode_p4": {
+        "bin": "nemoguardtopic8b_decode_p4_w8.bin"
+      },
+      "decode_p5": {
+        "bin": "nemoguardtopic8b_decode_p5_w8.bin"
+      },
+      "decode_p6": {
+        "bin": "nemoguardtopic8b_decode_p6_w8.bin"
+      },
+      "decode_p7": {
+        "bin": "nemoguardtopic8b_decode_p7_w8.bin"
+      },
+      "lmhead": {
+        "bin": "nemoguardtopic8b_lmhead_w8.bin"
+      }
+    },
+    "embed": "nemoguardtopic8b_embed_f16.bin",
+    "tokenizer": "tokenizer.json"
+  },
+  "plan": {
+    "steps": [
+      {
+        "host": "llama_generate_sharded",
+        "params": {
+          "decode_parts": "nemoguardtopic8b_decode_p0_w8,nemoguardtopic8b_decode_p1_w8,nemoguardtopic8b_decode_p2_w8,nemoguardtopic8b_decode_p3_w8,nemoguardtopic8b_decode_p4_w8,nemoguardtopic8b_decode_p5_w8,nemoguardtopic8b_decode_p6_w8,nemoguardtopic8b_decode_p7_w8",
+          "lmhead": "nemoguardtopic8b_lmhead_w8",
+          "max_new": 64,
+          "neg": -50000,
+          "rope_factor": 8.0
+        }
+      }
+    ]
+  }
+}

v81/nemoguardtopic8b_decode_p0_w8.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e6816dd61dc245734ebc762ea1a2b6729a2cd3a6dc1c34d5e823e27e5851bf12
+size 876814336

v81/nemoguardtopic8b_decode_p1_w8.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f95c20d1dc5672811860a4539a7bd5682e5ef5eb7790f4d77cb61675ac40892e
+size 876814336

v81/nemoguardtopic8b_decode_p2_w8.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ea5e79b49aaccc67836f8a32912e3a8f46cd2a61094f1bc8b4a4f01c6f03b4dc
+size 876814336

v81/nemoguardtopic8b_decode_p3_w8.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:390c481930d05350b3da7eafdd1bb845300d5468b8db309604eb9b7b39212af7
+size 876814336

v81/nemoguardtopic8b_decode_p4_w8.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:06bac3ddac7710e92151184e5825131122830e6aa0ac0fe23e88d8170e87d9ab
+size 876814336

v81/nemoguardtopic8b_decode_p5_w8.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f588253aff23393c7fad7688f10808a5685e7c204574e89bf10dfa76a30a40b8
+size 876814336

v81/nemoguardtopic8b_decode_p6_w8.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d306cd485be5be75c2af0ea678ca5f9b80d4c240f01c903cb3949f2857e947fa
+size 876814336

v81/nemoguardtopic8b_decode_p7_w8.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:55d0f6f096f38fccfbf278742959fe21a68b1d170d4d30907be3a2122edc64e7
+size 876826624

v81/nemoguardtopic8b_embed_f16.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e979c418bb40e497b73c6ad34d90be9a44c9f5373336cf49249c33f550f4036f
+size 1050673152

v81/nemoguardtopic8b_lmhead_w8.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:36c21e5d694e7f63d804f7d2e9915be5798e2e15f2bd42dff89c9bc50293d3bd
+size 527282176

v81/tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6b9e4e7fb171f92fd137b777cc2714bf87d11576700a1dcd7a399e7bbe39537b
+size 17209920