xpuenabler/OpenVLA-OFT-LIBERO-130-QNN-HTP

By xpuenabler · ↓ 0 · ♥ 0 · Updated 2026-03-02

What's in this repo

This page lists the files in xpuenabler/OpenVLA-OFT-LIBERO-130-QNN-HTP, including action_head.onnx, llm_backbone_v9.onnx, vision_encoder.onnx.

3 ONNX models (.onnx files)
1180 files, including 299 .bias, 166 .weight, 46 .scale_factor, 3 .sh, 3 .onnx, 3 .so
Last updated 2026-03-02

ONNX files

File	Size	Download
onnx/action_head.onnx	—	Download
onnx/llm_backbone_v9.onnx	—	Download
onnx/vision_encoder.onnx	—	Download

How to use

These ONNX files can be loaded with ONNX Runtime. To inspect a model graph, upload the file to the Netron viewer. Download links point to the original files on Hugging Face.

README

OpenVLA-OFT LIBERO — QNN HTP (FP16, aarch64)

This repository contains QNN HTP backend model artifacts for running OpenVLA-OFT on a Qualcomm Snapdragon device with the Hexagon Tensor Processor (HTP / DSP).

All three sub-models are converted from FP32 ONNX to FP16 QNN and compiled into aarch64 shared libraries (.so) for the aarch64-oe-linux-gcc11.2 target.

CPU (FP32) counterpart: xpuenabler/OpenVLA-OFT-LIBERO-130-QNN-CPU

Repository Structure

.
├── README.md
├── conversion/
│   ├── make_llm_v9.py              # Patch llm_backbone_v7.onnx → v9 (required before LLM conversion)
│   ├── convert_ve_htp.sh           # Vision Encoder: ONNX → QNN FP16 → libvision_encoder_sliced_htp.so
│   ├── convert_ah_htp.sh           # Action Head:   ONNX → QNN FP16 → libaction_head_htp.so
│   └── convert_llm_v9_htp.sh      # LLM Backbone:  ONNX → QNN FP16 → libllm_backbone_htp.so
│
├── onnx/
│   ├── vision_encoder.onnx         # VE ONNX (self-contained, 894 KB)
│   ├── action_head.onnx            # AH ONNX (self-contained, 4 KB)
│   ├── llm_backbone_v9.onnx        # LLM ONNX header (2.3 MB); requires external data files below
│   ├── embed_tokens.weight         # LLM external data (502 MB)
│   └── onnx__MatMul_*/             # LLM external data (~25 GB total, 289 files)
│
└── qnn_models/
    ├── libvision_encoder_sliced_htp.so   # VE compiled HTP library  (1.5 GB, aarch64)
    ├── libaction_head_htp.so             # AH compiled HTP library  (251 MB, aarch64)
    └── libllm_backbone_htp.so            # LLM compiled HTP library (13  GB, aarch64)

Prerequisites

Hardware / OS

Build host: Linux x86-64 (Ubuntu 20.04 tested)
Target device: Qualcomm SoC with HTP (Snapdragon 8 Gen 2 / 3 / Elite, etc.)

Software

Component	Version / Path
Qualcomm AI Runtime (QNN) SDK	2.43.0.260128
Python (QNN converter env)	3.10 (`miniforge3/envs/qnn`)
LLVM / clang++	17.0.6 (cross-compile to aarch64)
GCC aarch64 toolchain	11.2.1 (`aarch64-none-linux-gnu`)
ld.bfd (aarch64)	bundled with GCC toolchain

Python packages (`qnn` conda env)

onnx
onnxruntime
qti.aisw (from QNN SDK lib/python)

Required directories

QNN_ROOT=/path/to/qairt/2.43.0.260128
QENV=/path/to/miniforge3/envs/qnn
LLVM17=/path/to/llvm17/bin               # clang++ 17 with aarch64 target
AARCH64_PREFIX=/path/to/arm_gcc11/bin/aarch64-none-linux-gnu-
SYSROOT=/path/to/arm_gcc11/aarch64-none-linux-gnu/libc

Environment Setup

1. Set up QNN Python environment

# Activate the QNN conda env
conda activate qnn   # Python 3.10

# Add QNN Python bindings to PYTHONPATH
export QNN_ROOT=/path/to/qairt/2.43.0.260128
export DEPS=/path/to/qnn_py310_deps          # extra QNN runtime deps

export LD_LIBRARY_PATH=$QNN_ROOT/lib/x86_64-linux-clang:$LD_LIBRARY_PATH
export PYTHONPATH=$DEPS:$QNN_ROOT/lib/python
export PYTHONNOUSERSITE=1   # prevents user-site package conflicts
export TMPDIR=/tmp/qnn_tmp  # set to a large-capacity directory
mkdir -p $TMPDIR

2. Verify toolchain

# LLVM17 clang++ must support aarch64 target
clang++ --version          # → clang version 17.x
clang++ -print-targets | grep aarch64  # → aarch64

# aarch64-none-linux-gnu-g++ and ld.bfd must exist
aarch64-none-linux-gnu-g++ --version   # → gcc 11.2.1
aarch64-none-linux-gnu-ld.bfd --version

3. Prepare LLM ONNX (v9 patch)

The LLM backbone ONNX must be patched from the original v7 before conversion:

# Requires: llm_backbone_v7.onnx + external data files in the same directory
cd /path/to/onnx_models_fp32/
PYTHONNOUSERSITE=1 python3.10 make_llm_v9.py
# Produces: llm_backbone_v9.onnx

make_llm_v9.py is provided in conversion/.

Conversion

Edit the path variables at the top of each script to match your environment, then run:

Vision Encoder

bash conversion/convert_ve_htp.sh
# Output: $WORK/model_lib_aarch64/libs/aarch64-oe-linux-gcc11.2/libvision_encoder_sliced_htp.so
# Runtime: ~3 min

Key parameters:

Input: pixel_values shape 1,6,224,224
--float_bitwidth 16 (FP16 for HTP)

Action Head

bash conversion/convert_ah_htp.sh
# Output: $WORK/model_lib_aarch64/libs/aarch64-oe-linux-gcc11.2/libaction_head_htp.so
# Runtime: ~1 min

Key parameters:

Input: hidden_states shape 1,440,4096 (Note: ONNX model stores this as [1,440,4096], not [1,4096,440])
--float_bitwidth 16

LLM Backbone

bash conversion/convert_llm_v9_htp.sh
# Output: $WORK/model_lib_aarch64/libs/aarch64-oe-linux-gcc11.2/libllm_backbone_htp.so
# Runtime: qnn-onnx-converter ~5 min, compilation ~30 min (13 GB binary)

Key parameters:

Inputs: input_ids_no_stop 1,184, patch_features 1,256,4096, attention_mask 1,184
--float_bitwidth 16

Step-by-Step Conversion Detail

Each script performs the same three stages:

Stage 1 — ONNX → QNN source (qnn-onnx-converter)

$QENV/bin/python3.10 $QNN_ROOT/bin/x86_64-linux-clang/qnn-onnx-converter \
  --input_network <model>.onnx \
  --input_dim <name> <shape> \
  --output_path $WORK/<model> \
  --float_bitwidth 16

Outputs: <model> (C++ source, no extension) + <model>.bin (weight tarball).

Note: The converter produces the source file without a .cpp extension. The scripts copy it to <model>.cpp before compilation.

Stage 2 — Build directory setup

# Copy SDK helper files into jni/
SDK_JNI=$QNN_ROOT/share/QNN/converter/jni
cp $SDK_JNI/QnnModel.cpp $SDK_JNI/QnnModel.hpp ... $JNI/
cp $WORK/<model>.cpp $JNI/

# Copy + patch SDK Makefile
cp $QNN_ROOT/share/QNN/converter/Makefile.oe-linux-aarch64-gcc11.2 $BUILD/
# Patch: replace gnu++20 with c++11 + add -Wno-c99-designator
sed -i 's/-std=gnu++20 -D_LINUX_OE_SOURCE -fPIC -Wl,-lstdc++/-std=c++11 -Wno-c99-designator/' \
  $BUILD/Makefile.oe-linux-aarch64-gcc11.2
# LLM only: add -nostartfiles to LDFLAGS
sed -i 's/LDFLAGS += -shared -s/LDFLAGS += -shared -s -nostartfiles/' \
  $BUILD/Makefile.oe-linux-aarch64-gcc11.2

Stage 3 — Compile to aarch64 shared library

# VE / AH (standard linker)
CXX_CROSS="$LLVM17/clang++ --target=aarch64-linux-gnu \
           --sysroot=$SYSROOT --gcc-toolchain=$GCC_TOOLCHAIN"

# LLM (bfd linker required for 13 GB binary)
CXX_CROSS="$LLVM17/clang++ --target=aarch64-linux-gnu \
           --sysroot=$SYSROOT --gcc-toolchain=$GCC_TOOLCHAIN \
           -fuse-ld=${AARCH64_PREFIX}ld.bfd"

make -C $BUILD -f Makefile.oe-linux-aarch64-gcc11.2 \
  CXX="$CXX_CROSS" \
  TARGET_PREFIX=${AARCH64_PREFIX} \
  QNN_SDK_ROOT=$QNN_ROOT \
  QNN_MODEL_LIB_NAME=lib<model>_htp

Technical Notes

Why FP16 for HTP?

The Hexagon Tensor Processor natively accelerates FP16 operations. FP32 support on HTP is limited; using --float_bitwidth 16 is recommended for maximum HTP utilization.

`qnn-model-lib-generator` cannot be used for large models

qnn-model-lib-generator cannot compile the LLM backbone (13 GB binary) because it spawns a subprocess that cannot find the required cross-compiler toolchain environment variables (${TARGET_PREFIX}g++, ${SDKTARGETSYSROOT}). Manual Makefile-based compilation is used instead.

C++ standard: gnu++20 → c++11

The QNN SDK's Makefile.oe-linux-aarch64-gcc11.2 uses -std=gnu++20. However, QNN-generated C++ code mixes designated and non-designated initializers (a C99 extension not fully supported in C++20 strict mode). We patch the Makefile to use -std=c++11 -Wno-c99-designator.

aarch64 large model relocation issue (LLM only)

Compiling a 13 GB shared library with -mcmodel=large produces R_AARCH64_MOVW_UA relocations in .data.rel.ro that neither LLD nor gold can satisfy. The fix is to use the GNU BFD linker (aarch64-none-linux-gnu-ld.bfd) with -nostartfiles:

-fuse-ld=/path/to/aarch64-none-linux-gnu-ld.bfd -nostartfiles

This matches the approach used for the CPU x86-64 build where -fuse-ld=bfd -nostartfiles was needed.

PYTHONNOUSERSITE=1 is mandatory

Without PYTHONNOUSERSITE=1, a user-level numpy installation (Python 3.12 ABI) conflicts with the conda env's Python 3.10/3.11, causing an ImportError when the QNN converter tries to import onnx.

Model Information

Sub-model	ONNX Input	Precision	.so size
Vision Encoder	`pixel_values [1,6,224,224]`	FP16	1.5 GB
Action Head	`hidden_states [1,440,4096]`	FP16	251 MB
LLM Backbone	`input_ids_no_stop [1,184]`, `patch_features [1,256,4096]`, `attention_mask [1,184]`	FP16	13 GB

Base model: openvla/openvla-oft-libero QNN SDK: Qualcomm AI Runtime 2.43.0.260128 Target: aarch64-oe-linux-gcc11.2 (Snapdragon HTP)

Model info

onnx region:us

↓ 0 ♥ 0

Created: 2026-03-02

Updated: 2026-03-02

View on Hugging Face

Files (1180)

.gitattributes

README.md

conversion/convert_ah_htp.sh

conversion/convert_llm_v9_htp.sh

conversion/convert_ve_htp.sh

conversion/make_llm_v9.py

onnx/15b62db8-12fa-11f1-a779-d653e5e1fbe9

onnx/15b67e3e-12fa-11f1-afee-d653e5e1fbe9

onnx/15b67f0c-12fa-11f1-a90d-d653e5e1fbe9

onnx/15b67f93-12fa-11f1-8b88-d653e5e1fbe9

onnx/15b68001-12fa-11f1-acb7-d653e5e1fbe9

onnx/15b68064-12fa-11f1-a54b-d653e5e1fbe9

onnx/15b680c7-12fa-11f1-9e5f-d653e5e1fbe9

onnx/15b68138-12fa-11f1-a14d-d653e5e1fbe9

onnx/15b68196-12fa-11f1-ba59-d653e5e1fbe9

onnx/15b681ec-12fa-11f1-98d7-d653e5e1fbe9

onnx/15b68241-12fa-11f1-9959-d653e5e1fbe9

onnx/15b682a4-12fa-11f1-bc09-d653e5e1fbe9

onnx/15b68304-12fa-11f1-a3ff-d653e5e1fbe9

onnx/15b68363-12fa-11f1-ba33-d653e5e1fbe9

onnx/15b683be-12fa-11f1-8103-d653e5e1fbe9

onnx/15b68418-12fa-11f1-ac5c-d653e5e1fbe9

onnx/15b6846d-12fa-11f1-b71d-d653e5e1fbe9

onnx/15b684e7-12fa-11f1-81df-d653e5e1fbe9

onnx/15b6853e-12fa-11f1-8ede-d653e5e1fbe9

onnx/15b68596-12fa-11f1-9fd3-d653e5e1fbe9

onnx/15b685ea-12fa-11f1-9df5-d653e5e1fbe9

onnx/15b68649-12fa-11f1-ae89-d653e5e1fbe9

onnx/15b686a3-12fa-11f1-892c-d653e5e1fbe9

onnx/15b686fd-12fa-11f1-b789-d653e5e1fbe9

onnx/15b68751-12fa-11f1-899b-d653e5e1fbe9

onnx/15b6881a-12fa-11f1-9c3d-d653e5e1fbe9

onnx/15b68885-12fa-11f1-96d6-d653e5e1fbe9

onnx/15b688ef-12fa-11f1-9926-d653e5e1fbe9

onnx/15b68948-12fa-11f1-8a25-d653e5e1fbe9

onnx/15b6899b-12fa-11f1-8e81-d653e5e1fbe9

onnx/15b689ed-12fa-11f1-b92b-d653e5e1fbe9

onnx/15b68a3f-12fa-11f1-b69d-d653e5e1fbe9

onnx/15b68a94-12fa-11f1-ae7b-d653e5e1fbe9

onnx/15b68ae6-12fa-11f1-a03c-d653e5e1fbe9

onnx/15b68b3a-12fa-11f1-9edc-d653e5e1fbe9

onnx/15b68ba2-12fa-11f1-a053-d653e5e1fbe9

onnx/15b68bfa-12fa-11f1-bc16-d653e5e1fbe9

onnx/15b68c65-12fa-11f1-b211-d653e5e1fbe9

onnx/15b68cbc-12fa-11f1-b8fe-d653e5e1fbe9

onnx/15b68d15-12fa-11f1-85b3-d653e5e1fbe9

onnx/15b68d6e-12fa-11f1-9df8-d653e5e1fbe9

onnx/15b68dc4-12fa-11f1-be8c-d653e5e1fbe9

onnx/15b68e19-12fa-11f1-a8ee-d653e5e1fbe9

onnx/15b68e6e-12fa-11f1-99fa-d653e5e1fbe9

onnx/15b68ec5-12fa-11f1-9800-d653e5e1fbe9

onnx/15b68f1b-12fa-11f1-84f6-d653e5e1fbe9

onnx/15b68f6d-12fa-11f1-8aa0-d653e5e1fbe9

onnx/15b68fbf-12fa-11f1-b890-d653e5e1fbe9

onnx/15b69012-12fa-11f1-9040-d653e5e1fbe9

onnx/15b6907b-12fa-11f1-965c-d653e5e1fbe9

onnx/15b690d1-12fa-11f1-9695-d653e5e1fbe9

onnx/15b6912c-12fa-11f1-81fc-d653e5e1fbe9

onnx/15b69185-12fa-11f1-a2fa-d653e5e1fbe9

onnx/15b691de-12fa-11f1-b07b-d653e5e1fbe9

onnx/15b69254-12fa-11f1-a550-d653e5e1fbe9

onnx/15b692ac-12fa-11f1-9084-d653e5e1fbe9

onnx/15b69305-12fa-11f1-927b-d653e5e1fbe9

onnx/15b6935b-12fa-11f1-b42b-d653e5e1fbe9

onnx/15b693b8-12fa-11f1-a484-d653e5e1fbe9

onnx/15b6940a-12fa-11f1-a1a7-d653e5e1fbe9

onnx/15b69460-12fa-11f1-a73a-d653e5e1fbe9

onnx/15b694af-12fa-11f1-a003-d653e5e1fbe9

onnx/15b69502-12fa-11f1-b9dc-d653e5e1fbe9

onnx/15b69575-12fa-11f1-a454-d653e5e1fbe9

onnx/15b695cb-12fa-11f1-8300-d653e5e1fbe9

onnx/15b6961e-12fa-11f1-8bc4-d653e5e1fbe9

onnx/15b6966f-12fa-11f1-a16e-d653e5e1fbe9

onnx/15b696c1-12fa-11f1-8a35-d653e5e1fbe9

onnx/15b69718-12fa-11f1-9036-d653e5e1fbe9

onnx/15b6976a-12fa-11f1-8ef7-d653e5e1fbe9

onnx/15b697c0-12fa-11f1-b05d-d653e5e1fbe9

onnx/15b69813-12fa-11f1-a725-d653e5e1fbe9

onnx/15b69863-12fa-11f1-8838-d653e5e1fbe9

onnx/15b698b5-12fa-11f1-ac83-d653e5e1fbe9

onnx/15b69906-12fa-11f1-8b51-d653e5e1fbe9

onnx/15b6995b-12fa-11f1-886b-d653e5e1fbe9

onnx/15b699af-12fa-11f1-964f-d653e5e1fbe9

onnx/15b69a16-12fa-11f1-afaa-d653e5e1fbe9

onnx/15b69a68-12fa-11f1-ad24-d653e5e1fbe9

onnx/15b69ac7-12fa-11f1-a322-d653e5e1fbe9

onnx/15b69b1a-12fa-11f1-b585-d653e5e1fbe9

onnx/15b69b6d-12fa-11f1-a744-d653e5e1fbe9

onnx/15b69bc4-12fa-11f1-b951-d653e5e1fbe9

onnx/15b69d1e-12fa-11f1-91f8-d653e5e1fbe9

onnx/15b69d82-12fa-11f1-9835-d653e5e1fbe9

onnx/15b69dd5-12fa-11f1-ad66-d653e5e1fbe9

onnx/15b69e27-12fa-11f1-8ebf-d653e5e1fbe9

onnx/15b69e7c-12fa-11f1-9576-d653e5e1fbe9

onnx/15b69ede-12fa-11f1-a8e6-d653e5e1fbe9

onnx/15b69f4d-12fa-11f1-9c8d-d653e5e1fbe9

onnx/15b69fb1-12fa-11f1-bbbd-d653e5e1fbe9

onnx/15b6a038-12fa-11f1-a88f-d653e5e1fbe9

onnx/15b6a096-12fa-11f1-ac37-d653e5e1fbe9

onnx/15b6a0ef-12fa-11f1-916b-d653e5e1fbe9

onnx/15b6a146-12fa-11f1-ad8a-d653e5e1fbe9

onnx/15b6a1ab-12fa-11f1-8da1-d653e5e1fbe9

onnx/15b6a200-12fa-11f1-850d-d653e5e1fbe9

onnx/15b6a259-12fa-11f1-9daf-d653e5e1fbe9

onnx/15b6a2b5-12fa-11f1-8c93-d653e5e1fbe9

onnx/15b6a310-12fa-11f1-8a93-d653e5e1fbe9

onnx/15b6a365-12fa-11f1-9cae-d653e5e1fbe9

onnx/15b6a3bc-12fa-11f1-a8ab-d653e5e1fbe9

onnx/15b6a40f-12fa-11f1-b48f-d653e5e1fbe9

onnx/15b6a465-12fa-11f1-a9f5-d653e5e1fbe9

onnx/15b6a4bb-12fa-11f1-9b33-d653e5e1fbe9

onnx/15b6a529-12fa-11f1-843b-d653e5e1fbe9

onnx/15b6a580-12fa-11f1-ba31-d653e5e1fbe9

onnx/15b6a5d9-12fa-11f1-b69a-d653e5e1fbe9

onnx/15b6a62b-12fa-11f1-b913-d653e5e1fbe9

onnx/15b6a67b-12fa-11f1-9568-d653e5e1fbe9

onnx/15b6a6ca-12fa-11f1-9d00-d653e5e1fbe9

onnx/15b6a735-12fa-11f1-9dee-d653e5e1fbe9

onnx/15b6a788-12fa-11f1-8f10-d653e5e1fbe9

onnx/15b6a7d7-12fa-11f1-aa60-d653e5e1fbe9

onnx/15b6a843-12fa-11f1-af3a-d653e5e1fbe9

onnx/15b6a8a1-12fa-11f1-aa56-d653e5e1fbe9

onnx/15b6a8f6-12fa-11f1-af0c-d653e5e1fbe9

onnx/15b6a950-12fa-11f1-a612-d653e5e1fbe9

onnx/15b6a9a5-12fa-11f1-bd7c-d653e5e1fbe9

onnx/15b6aa08-12fa-11f1-a834-d653e5e1fbe9

onnx/15b6aa5e-12fa-11f1-8bfd-d653e5e1fbe9

onnx/15b6aabd-12fa-11f1-9025-d653e5e1fbe9

onnx/15b6ab11-12fa-11f1-bb25-d653e5e1fbe9

onnx/15b6ab66-12fa-11f1-9c30-d653e5e1fbe9

onnx/15b6abb7-12fa-11f1-a5e2-d653e5e1fbe9

onnx/15b6ac09-12fa-11f1-acc4-d653e5e1fbe9

onnx/15b6ac73-12fa-11f1-a7f9-d653e5e1fbe9

onnx/15b6acda-12fa-11f1-9bbd-d653e5e1fbe9

onnx/15b6ad2f-12fa-11f1-9ac6-d653e5e1fbe9

onnx/15b6ad86-12fa-11f1-a7ad-d653e5e1fbe9

onnx/15b6addb-12fa-11f1-ba1a-d653e5e1fbe9

onnx/15b6ae30-12fa-11f1-b2e8-d653e5e1fbe9

onnx/15b6af6d-12fa-11f1-b3cd-d653e5e1fbe9

onnx/15b6b00b-12fa-11f1-b0aa-d653e5e1fbe9

onnx/15b6b06c-12fa-11f1-8e2b-d653e5e1fbe9

onnx/15b6b0bf-12fa-11f1-b136-d653e5e1fbe9

onnx/15b6b111-12fa-11f1-82fd-d653e5e1fbe9

onnx/15b6b167-12fa-11f1-947e-d653e5e1fbe9

onnx/15b6b1bc-12fa-11f1-ad86-d653e5e1fbe9

onnx/15b6b220-12fa-11f1-9d75-d653e5e1fbe9

onnx/15b6b271-12fa-11f1-85a5-d653e5e1fbe9

onnx/15b6b2c2-12fa-11f1-8141-d653e5e1fbe9

onnx/15b6b327-12fa-11f1-afbb-d653e5e1fbe9

onnx/15b6b37f-12fa-11f1-82cf-d653e5e1fbe9

onnx/15b6b3d4-12fa-11f1-ab4b-d653e5e1fbe9

onnx/15b6b427-12fa-11f1-b79a-d653e5e1fbe9

onnx/15b6b480-12fa-11f1-872e-d653e5e1fbe9

onnx/15b6b4e1-12fa-11f1-850f-d653e5e1fbe9

onnx/15b6b536-12fa-11f1-9c8a-d653e5e1fbe9

onnx/15b6b589-12fa-11f1-a623-d653e5e1fbe9

onnx/15b6b5dc-12fa-11f1-90bd-d653e5e1fbe9

onnx/15b6b634-12fa-11f1-be86-d653e5e1fbe9

onnx/15b6b6b1-12fa-11f1-854e-d653e5e1fbe9

onnx/15b6b703-12fa-11f1-bbf0-d653e5e1fbe9

onnx/15b6b756-12fa-11f1-91de-d653e5e1fbe9

onnx/15b6b7a4-12fa-11f1-951c-d653e5e1fbe9

onnx/15b6b7f6-12fa-11f1-b556-d653e5e1fbe9

onnx/15b6b846-12fa-11f1-82ad-d653e5e1fbe9

onnx/15b6b8a4-12fa-11f1-8509-d653e5e1fbe9

onnx/15b6b8f5-12fa-11f1-8742-d653e5e1fbe9

onnx/15b6b948-12fa-11f1-b2d2-d653e5e1fbe9

onnx/15b6b9b0-12fa-11f1-bcef-d653e5e1fbe9

onnx/15b6ba13-12fa-11f1-bd38-d653e5e1fbe9

onnx/15b6ba6b-12fa-11f1-ba56-d653e5e1fbe9

onnx/15b6bac6-12fa-11f1-8099-d653e5e1fbe9

onnx/15b6bb35-12fa-11f1-87d7-d653e5e1fbe9

onnx/15b6bb8d-12fa-11f1-8fc8-d653e5e1fbe9

onnx/15b6bbe3-12fa-11f1-add3-d653e5e1fbe9

onnx/15b6bc38-12fa-11f1-9250-d653e5e1fbe9

onnx/15b6bc90-12fa-11f1-8d9a-d653e5e1fbe9

onnx/15b6bceb-12fa-11f1-812a-d653e5e1fbe9

onnx/15b6bd48-12fa-11f1-ba7a-d653e5e1fbe9

onnx/15b6bd9c-12fa-11f1-9b2c-d653e5e1fbe9

onnx/15b6bded-12fa-11f1-a9cd-d653e5e1fbe9

onnx/15b6be55-12fa-11f1-bb55-d653e5e1fbe9

onnx/15b6beb9-12fa-11f1-972a-d653e5e1fbe9

onnx/15b6bf0f-12fa-11f1-a832-d653e5e1fbe9

onnx/15b6bf61-12fa-11f1-b3b9-d653e5e1fbe9

onnx/15b6bfb8-12fa-11f1-accc-d653e5e1fbe9

onnx/15b6c00c-12fa-11f1-a19f-d653e5e1fbe9

onnx/15b6c060-12fa-11f1-ae9a-d653e5e1fbe9

onnx/15b6c0b8-12fa-11f1-9e3d-d653e5e1fbe9

onnx/15b6c10c-12fa-11f1-9476-d653e5e1fbe9

onnx/15b6c23a-12fa-11f1-ae58-d653e5e1fbe9

onnx/15b6c297-12fa-11f1-8d74-d653e5e1fbe9

onnx/15b6c2ee-12fa-11f1-9b25-d653e5e1fbe9

onnx/15b6c342-12fa-11f1-b436-d653e5e1fbe9

onnx/15b6c397-12fa-11f1-8ab1-d653e5e1fbe9

onnx/15b6c3e8-12fa-11f1-bc4a-d653e5e1fbe9

onnx/15b6c44b-12fa-11f1-a48b-d653e5e1fbe9

onnx/15b6c4bc-12fa-11f1-bc08-d653e5e1fbe9

onnx/15b6c512-12fa-11f1-ac65-d653e5e1fbe9

onnx/15b6c578-12fa-11f1-9f87-d653e5e1fbe9

onnx/15b6c5cb-12fa-11f1-a326-d653e5e1fbe9

onnx/15b6c61f-12fa-11f1-be8a-d653e5e1fbe9

onnx/15b6c675-12fa-11f1-a691-d653e5e1fbe9

onnx/15b6c6ca-12fa-11f1-8f82-d653e5e1fbe9

onnx/15b6c71b-12fa-11f1-86d6-d653e5e1fbe9

onnx/15b6c771-12fa-11f1-9160-d653e5e1fbe9

onnx/17ec313f-12fa-11f1-bd77-d653e5e1fbe9

onnx/35b3799e-12fa-11f1-b233-d653e5e1fbe9

onnx/35b3cd8d-12fa-11f1-b891-d653e5e1fbe9

onnx/35b3ce73-12fa-11f1-ae47-d653e5e1fbe9

onnx/35b3cf10-12fa-11f1-a7cb-d653e5e1fbe9

onnx/35b3cf8b-12fa-11f1-9192-d653e5e1fbe9

onnx/35b3d009-12fa-11f1-a5d0-d653e5e1fbe9

onnx/35b3d071-12fa-11f1-97cd-d653e5e1fbe9

onnx/35b3d0da-12fa-11f1-b295-d653e5e1fbe9

onnx/35b3d142-12fa-11f1-a32c-d653e5e1fbe9

onnx/35b3d1a1-12fa-11f1-b27d-d653e5e1fbe9

onnx/35b3d216-12fa-11f1-8fad-d653e5e1fbe9

onnx/35b3d28a-12fa-11f1-8758-d653e5e1fbe9

onnx/35b3d2f1-12fa-11f1-91f1-d653e5e1fbe9

onnx/35b3d34d-12fa-11f1-8ae8-d653e5e1fbe9

onnx/35b3d3ab-12fa-11f1-a4cd-d653e5e1fbe9

onnx/35b3d406-12fa-11f1-864e-d653e5e1fbe9

onnx/35b3d469-12fa-11f1-a72a-d653e5e1fbe9

onnx/35b3d4c5-12fa-11f1-a8c5-d653e5e1fbe9

onnx/35b3d528-12fa-11f1-8573-d653e5e1fbe9

onnx/35b3d58f-12fa-11f1-b459-d653e5e1fbe9

onnx/35b3d5f3-12fa-11f1-b9be-d653e5e1fbe9

onnx/35b3d659-12fa-11f1-abf0-d653e5e1fbe9

onnx/35b3d6b5-12fa-11f1-92f6-d653e5e1fbe9

onnx/35b3d70b-12fa-11f1-928e-d653e5e1fbe9

onnx/35b3d761-12fa-11f1-8c0a-d653e5e1fbe9

onnx/35b3d7cb-12fa-11f1-bd17-d653e5e1fbe9

onnx/35b3d823-12fa-11f1-946d-d653e5e1fbe9

onnx/35b3d87f-12fa-11f1-b6c8-d653e5e1fbe9

onnx/35b3d8d9-12fa-11f1-b4ab-d653e5e1fbe9

onnx/35b3d941-12fa-11f1-bcc1-d653e5e1fbe9

onnx/35b3d99d-12fa-11f1-a598-d653e5e1fbe9

onnx/35b3d9f7-12fa-11f1-acab-d653e5e1fbe9

onnx/35b3da52-12fa-11f1-919f-d653e5e1fbe9

onnx/35b3daa9-12fa-11f1-ad66-d653e5e1fbe9

onnx/35b3db05-12fa-11f1-9250-d653e5e1fbe9

onnx/35b3db5f-12fa-11f1-a5b6-d653e5e1fbe9

onnx/35b3dbb4-12fa-11f1-a5c3-d653e5e1fbe9

onnx/35b3dc09-12fa-11f1-8739-d653e5e1fbe9

onnx/35b3dc62-12fa-11f1-ac78-d653e5e1fbe9

onnx/35b3dcc5-12fa-11f1-9aee-d653e5e1fbe9

onnx/35b3dd1b-12fa-11f1-95d9-d653e5e1fbe9

onnx/35b3dd74-12fa-11f1-9282-d653e5e1fbe9

onnx/35b3ddcf-12fa-11f1-b021-d653e5e1fbe9

onnx/35b3de29-12fa-11f1-88a5-d653e5e1fbe9

onnx/35b3de82-12fa-11f1-8cd0-d653e5e1fbe9

onnx/35b3dee6-12fa-11f1-96a2-d653e5e1fbe9

onnx/35b3df43-12fa-11f1-9b95-d653e5e1fbe9

onnx/35b3df98-12fa-11f1-a7a5-d653e5e1fbe9

onnx/35b3dff6-12fa-11f1-8aff-d653e5e1fbe9

onnx/35b3e604-12fa-11f1-a939-d653e5e1fbe9

onnx/35b3e66f-12fa-11f1-a03f-d653e5e1fbe9

onnx/35b3e6d3-12fa-11f1-be10-d653e5e1fbe9

onnx/35b3e729-12fa-11f1-94a0-d653e5e1fbe9

onnx/35b3e77e-12fa-11f1-bcbd-d653e5e1fbe9

onnx/35b3e7f1-12fa-11f1-b50f-d653e5e1fbe9

onnx/35b3e84d-12fa-11f1-92c7-d653e5e1fbe9

onnx/35b3e8a0-12fa-11f1-95ca-d653e5e1fbe9

onnx/35b3e8f6-12fa-11f1-abe6-d653e5e1fbe9

onnx/35b3e952-12fa-11f1-8f73-d653e5e1fbe9

onnx/35b3e9aa-12fa-11f1-86e8-d653e5e1fbe9

onnx/35b3ea01-12fa-11f1-b978-d653e5e1fbe9

onnx/35b3ea55-12fa-11f1-b478-d653e5e1fbe9

onnx/35b3eaaa-12fa-11f1-b291-d653e5e1fbe9

onnx/35b3eafb-12fa-11f1-bc14-d653e5e1fbe9

onnx/35b3eb4f-12fa-11f1-bbe5-d653e5e1fbe9

onnx/35b3eba0-12fa-11f1-aa1c-d653e5e1fbe9

onnx/35b3ec56-12fa-11f1-b69c-d653e5e1fbe9

onnx/35b3eccd-12fa-11f1-9127-d653e5e1fbe9

onnx/35b3ed21-12fa-11f1-8dcc-d653e5e1fbe9

onnx/35b3ed84-12fa-11f1-93c5-d653e5e1fbe9

onnx/35b3ede1-12fa-11f1-8293-d653e5e1fbe9

onnx/35b3ee3b-12fa-11f1-8e19-d653e5e1fbe9

onnx/35b3eea2-12fa-11f1-8dda-d653e5e1fbe9

onnx/35b3ef02-12fa-11f1-9462-d653e5e1fbe9

onnx/35b3ef5e-12fa-11f1-be6a-d653e5e1fbe9

onnx/35b3efbb-12fa-11f1-b714-d653e5e1fbe9

onnx/35b3f00f-12fa-11f1-944f-d653e5e1fbe9

onnx/35b3f065-12fa-11f1-926a-d653e5e1fbe9

onnx/35b3f0bd-12fa-11f1-aed8-d653e5e1fbe9

onnx/35b3f111-12fa-11f1-b941-d653e5e1fbe9

onnx/35b3f166-12fa-11f1-ba67-d653e5e1fbe9

onnx/35b3f1c5-12fa-11f1-b15c-d653e5e1fbe9

onnx/35b3f219-12fa-11f1-8095-d653e5e1fbe9

onnx/35b3f277-12fa-11f1-921d-d653e5e1fbe9

onnx/35b3f2d2-12fa-11f1-82b1-d653e5e1fbe9

onnx/35b3f32d-12fa-11f1-b440-d653e5e1fbe9

onnx/35b3f382-12fa-11f1-a0fc-d653e5e1fbe9

onnx/35b3f3d6-12fa-11f1-a89c-d653e5e1fbe9

onnx/35b3f42a-12fa-11f1-9985-d653e5e1fbe9

onnx/35b3f488-12fa-11f1-8ae6-d653e5e1fbe9

onnx/35b3f4e6-12fa-11f1-bd53-d653e5e1fbe9

onnx/35b3f53c-12fa-11f1-80e1-d653e5e1fbe9

onnx/35b3f594-12fa-11f1-b04a-d653e5e1fbe9

onnx/35b3f5f2-12fa-11f1-b76a-d653e5e1fbe9

onnx/35b3f643-12fa-11f1-bfe0-d653e5e1fbe9

onnx/35b3f699-12fa-11f1-b165-d653e5e1fbe9

onnx/35b3f6ee-12fa-11f1-8f0b-d653e5e1fbe9

onnx/35b3f767-12fa-11f1-b415-d653e5e1fbe9

onnx/35b3f7be-12fa-11f1-ae04-d653e5e1fbe9

onnx/35b3f81f-12fa-11f1-a716-d653e5e1fbe9

onnx/35b3f879-12fa-11f1-9705-d653e5e1fbe9

onnx/35b3f8d0-12fa-11f1-8688-d653e5e1fbe9

onnx/35b3f925-12fa-11f1-b83b-d653e5e1fbe9

onnx/35b3f97a-12fa-11f1-807f-d653e5e1fbe9

onnx/35b3f9d1-12fa-11f1-9a1f-d653e5e1fbe9

onnx/35b3fa21-12fa-11f1-8116-d653e5e1fbe9

onnx/35b3fa76-12fa-11f1-9e86-d653e5e1fbe9

onnx/35b3fac4-12fa-11f1-a338-d653e5e1fbe9

onnx/35b3fb18-12fa-11f1-a6e8-d653e5e1fbe9

onnx/35b3fb72-12fa-11f1-9007-d653e5e1fbe9

onnx/35b3fbe4-12fa-11f1-b5e5-d653e5e1fbe9

onnx/35b3fc46-12fa-11f1-a160-d653e5e1fbe9

onnx/35b3fca0-12fa-11f1-a0d0-d653e5e1fbe9

onnx/35b3fd06-12fa-11f1-a744-d653e5e1fbe9

onnx/35b3fd59-12fa-11f1-be21-d653e5e1fbe9

onnx/35b3fdab-12fa-11f1-9ad4-d653e5e1fbe9

onnx/35b3fe04-12fa-11f1-8c3b-d653e5e1fbe9

onnx/35b3fe5e-12fa-11f1-a6e1-d653e5e1fbe9

onnx/35b3feb4-12fa-11f1-99c5-d653e5e1fbe9

onnx/35b3ff08-12fa-11f1-9e79-d653e5e1fbe9

onnx/35b3ff71-12fa-11f1-aa71-d653e5e1fbe9

onnx/35b3ffd2-12fa-11f1-a3bd-d653e5e1fbe9

onnx/35b40024-12fa-11f1-a012-d653e5e1fbe9

onnx/35b40087-12fa-11f1-9a49-d653e5e1fbe9

onnx/35b400dc-12fa-11f1-b4d0-d653e5e1fbe9

onnx/35b4012e-12fa-11f1-8314-d653e5e1fbe9

onnx/35b4017e-12fa-11f1-a0c7-d653e5e1fbe9

onnx/35b401d6-12fa-11f1-89dc-d653e5e1fbe9

onnx/35b4022a-12fa-11f1-a3d8-d653e5e1fbe9

onnx/35b4028e-12fa-11f1-9b5e-d653e5e1fbe9

onnx/35b402e2-12fa-11f1-ae15-d653e5e1fbe9

onnx/35b40330-12fa-11f1-9b5f-d653e5e1fbe9

onnx/35b40381-12fa-11f1-926c-d653e5e1fbe9

onnx/35b403d5-12fa-11f1-9124-d653e5e1fbe9

onnx/35b4042a-12fa-11f1-b193-d653e5e1fbe9

onnx/35b40482-12fa-11f1-8490-d653e5e1fbe9

onnx/35b404d5-12fa-11f1-a75c-d653e5e1fbe9

onnx/35b40532-12fa-11f1-8269-d653e5e1fbe9

onnx/35b40585-12fa-11f1-a186-d653e5e1fbe9

onnx/35b405d5-12fa-11f1-99d8-d653e5e1fbe9

onnx/35b4062b-12fa-11f1-8813-d653e5e1fbe9

onnx/35b4067b-12fa-11f1-8ac9-d653e5e1fbe9

onnx/35b406d1-12fa-11f1-b045-d653e5e1fbe9

onnx/35b40725-12fa-11f1-9656-d653e5e1fbe9

onnx/35b40776-12fa-11f1-9019-d653e5e1fbe9

onnx/35b407f7-12fa-11f1-abd0-d653e5e1fbe9

onnx/35b4084f-12fa-11f1-85eb-d653e5e1fbe9

onnx/35b408a2-12fa-11f1-b9d6-d653e5e1fbe9

onnx/35b408f8-12fa-11f1-9d65-d653e5e1fbe9

onnx/35b40a36-12fa-11f1-971e-d653e5e1fbe9

onnx/35b40a91-12fa-11f1-9653-d653e5e1fbe9

onnx/35b40ae7-12fa-11f1-9a35-d653e5e1fbe9

onnx/35b40b3e-12fa-11f1-88d4-d653e5e1fbe9

onnx/35b40b8f-12fa-11f1-9d60-d653e5e1fbe9

onnx/35b40bdf-12fa-11f1-bb1d-d653e5e1fbe9

onnx/35b40c32-12fa-11f1-8577-d653e5e1fbe9

onnx/35b40c81-12fa-11f1-b407-d653e5e1fbe9

onnx/35b40cd0-12fa-11f1-b298-d653e5e1fbe9

onnx/35b40d21-12fa-11f1-8336-d653e5e1fbe9

onnx/35b40d77-12fa-11f1-a8a9-d653e5e1fbe9

onnx/35b40dc8-12fa-11f1-9f54-d653e5e1fbe9

onnx/35b40e2c-12fa-11f1-a7af-d653e5e1fbe9

onnx/35b40e7f-12fa-11f1-be68-d653e5e1fbe9

onnx/35b40ed6-12fa-11f1-901b-d653e5e1fbe9

onnx/35b40f2c-12fa-11f1-a603-d653e5e1fbe9

onnx/35b40f7e-12fa-11f1-aa93-d653e5e1fbe9

onnx/35b40fdd-12fa-11f1-b98c-d653e5e1fbe9

onnx/35b4102f-12fa-11f1-9ba5-d653e5e1fbe9

onnx/35b4107f-12fa-11f1-ae4b-d653e5e1fbe9

onnx/35b410d2-12fa-11f1-a65d-d653e5e1fbe9

onnx/35b41190-12fa-11f1-b1dd-d653e5e1fbe9

onnx/35b411fc-12fa-11f1-a7c5-d653e5e1fbe9

onnx/35b4124f-12fa-11f1-a950-d653e5e1fbe9

onnx/35b412a1-12fa-11f1-877b-d653e5e1fbe9

onnx/35b4134b-12fa-11f1-bef8-d653e5e1fbe9

onnx/35b413b5-12fa-11f1-b65c-d653e5e1fbe9

onnx/35b41408-12fa-11f1-a8ee-d653e5e1fbe9

onnx/35b41486-12fa-11f1-b8ff-d653e5e1fbe9

onnx/35b414de-12fa-11f1-82d7-d653e5e1fbe9

onnx/35b41532-12fa-11f1-9b47-d653e5e1fbe9

onnx/35b41599-12fa-11f1-b6b1-d653e5e1fbe9

onnx/35b415f4-12fa-11f1-b987-d653e5e1fbe9

onnx/35b41658-12fa-11f1-b668-d653e5e1fbe9

onnx/35b416a9-12fa-11f1-95be-d653e5e1fbe9

onnx/35b416fa-12fa-11f1-82a2-d653e5e1fbe9

onnx/35b4174f-12fa-11f1-b6b0-d653e5e1fbe9

onnx/35b4179f-12fa-11f1-8e64-d653e5e1fbe9

onnx/35b417f2-12fa-11f1-973a-d653e5e1fbe9

onnx/35b41844-12fa-11f1-9855-d653e5e1fbe9

onnx/35b41899-12fa-11f1-9326-d653e5e1fbe9

onnx/35b418ef-12fa-11f1-a8df-d653e5e1fbe9

onnx/35b41946-12fa-11f1-aef5-d653e5e1fbe9

onnx/35b41997-12fa-11f1-accb-d653e5e1fbe9

onnx/35b41a00-12fa-11f1-b10a-d653e5e1fbe9

onnx/35b41a5f-12fa-11f1-9968-d653e5e1fbe9

onnx/35b41ab1-12fa-11f1-b5d9-d653e5e1fbe9

onnx/35b41b17-12fa-11f1-9eba-d653e5e1fbe9

onnx/35b41b6e-12fa-11f1-8c04-d653e5e1fbe9

onnx/35b41bbb-12fa-11f1-a5a1-d653e5e1fbe9

onnx/35b41c0e-12fa-11f1-9207-d653e5e1fbe9

onnx/35b41c61-12fa-11f1-b96e-d653e5e1fbe9

onnx/35b41cb3-12fa-11f1-b458-d653e5e1fbe9

onnx/35b41d07-12fa-11f1-8c35-d653e5e1fbe9

onnx/35b41d5e-12fa-11f1-86f8-d653e5e1fbe9

onnx/35b41db2-12fa-11f1-a2c5-d653e5e1fbe9

onnx/35b41e05-12fa-11f1-9c02-d653e5e1fbe9

onnx/35b41e56-12fa-11f1-8c57-d653e5e1fbe9

onnx/35b41eaa-12fa-11f1-b8c9-d653e5e1fbe9

onnx/35b41f11-12fa-11f1-82b7-d653e5e1fbe9

onnx/35b41f7a-12fa-11f1-bc32-d653e5e1fbe9

onnx/35b41fce-12fa-11f1-8725-d653e5e1fbe9

onnx/35b4201f-12fa-11f1-8583-d653e5e1fbe9

onnx/35b4206c-12fa-11f1-b5db-d653e5e1fbe9

onnx/35b420c0-12fa-11f1-9b64-d653e5e1fbe9

onnx/35b42111-12fa-11f1-af06-d653e5e1fbe9

onnx/35b42160-12fa-11f1-b180-d653e5e1fbe9

onnx/35b421ae-12fa-11f1-b35d-d653e5e1fbe9

onnx/35b421fd-12fa-11f1-8f83-d653e5e1fbe9

onnx/35b42249-12fa-11f1-a13d-d653e5e1fbe9

onnx/35b42297-12fa-11f1-8bba-d653e5e1fbe9

onnx/35b422e6-12fa-11f1-be14-d653e5e1fbe9

onnx/35b4233e-12fa-11f1-bea1-d653e5e1fbe9

onnx/35b423b8-12fa-11f1-9b78-d653e5e1fbe9

onnx/35b42410-12fa-11f1-bb80-d653e5e1fbe9

onnx/_llama_model_Constant_1_attr__value

onnx/_llama_model_Constant_attr__value

onnx/action_head.onnx ONNX

onnx/bin_centers

onnx/embed_tokens.weight

onnx/llama_model.layers.0.input_layernorm.weight

onnx/llama_model.layers.0.post_attention_layernorm.weight

onnx/llama_model.layers.1.input_layernorm.weight

onnx/llama_model.layers.1.post_attention_layernorm.weight

onnx/llama_model.layers.10.input_layernorm.weight

onnx/llama_model.layers.10.post_attention_layernorm.weight

onnx/llama_model.layers.11.input_layernorm.weight

onnx/llama_model.layers.11.post_attention_layernorm.weight

onnx/llama_model.layers.12.input_layernorm.weight

onnx/llama_model.layers.12.post_attention_layernorm.weight

onnx/llama_model.layers.13.input_layernorm.weight

onnx/llama_model.layers.13.post_attention_layernorm.weight

onnx/llama_model.layers.14.input_layernorm.weight

onnx/llama_model.layers.14.post_attention_layernorm.weight

onnx/llama_model.layers.15.input_layernorm.weight

onnx/llama_model.layers.15.post_attention_layernorm.weight

onnx/llama_model.layers.16.input_layernorm.weight

onnx/llama_model.layers.16.post_attention_layernorm.weight

onnx/llama_model.layers.17.input_layernorm.weight

onnx/llama_model.layers.17.post_attention_layernorm.weight

onnx/llama_model.layers.18.input_layernorm.weight

onnx/llama_model.layers.18.post_attention_layernorm.weight

onnx/llama_model.layers.19.input_layernorm.weight

onnx/llama_model.layers.19.post_attention_layernorm.weight

onnx/llama_model.layers.2.input_layernorm.weight

onnx/llama_model.layers.2.post_attention_layernorm.weight

onnx/llama_model.layers.20.input_layernorm.weight

onnx/llama_model.layers.20.post_attention_layernorm.weight

onnx/llama_model.layers.21.input_layernorm.weight

onnx/llama_model.layers.21.post_attention_layernorm.weight

onnx/llama_model.layers.22.input_layernorm.weight

onnx/llama_model.layers.22.post_attention_layernorm.weight

onnx/llama_model.layers.23.input_layernorm.weight

onnx/llama_model.layers.23.post_attention_layernorm.weight

onnx/llama_model.layers.24.input_layernorm.weight

onnx/llama_model.layers.24.post_attention_layernorm.weight

onnx/llama_model.layers.25.input_layernorm.weight

onnx/llama_model.layers.25.post_attention_layernorm.weight

onnx/llama_model.layers.26.input_layernorm.weight

onnx/llama_model.layers.26.post_attention_layernorm.weight

onnx/llama_model.layers.27.input_layernorm.weight

onnx/llama_model.layers.27.post_attention_layernorm.weight

onnx/llama_model.layers.28.input_layernorm.weight

onnx/llama_model.layers.28.post_attention_layernorm.weight

onnx/llama_model.layers.29.input_layernorm.weight

onnx/llama_model.layers.29.post_attention_layernorm.weight

onnx/llama_model.layers.3.input_layernorm.weight

onnx/llama_model.layers.3.post_attention_layernorm.weight

onnx/llama_model.layers.30.input_layernorm.weight

onnx/llama_model.layers.30.post_attention_layernorm.weight

onnx/llama_model.layers.31.input_layernorm.weight

onnx/llama_model.layers.31.post_attention_layernorm.weight

onnx/llama_model.layers.4.input_layernorm.weight

onnx/llama_model.layers.4.post_attention_layernorm.weight

onnx/llama_model.layers.5.input_layernorm.weight

onnx/llama_model.layers.5.post_attention_layernorm.weight

onnx/llama_model.layers.6.input_layernorm.weight

onnx/llama_model.layers.6.post_attention_layernorm.weight

onnx/llama_model.layers.7.input_layernorm.weight

onnx/llama_model.layers.7.post_attention_layernorm.weight

onnx/llama_model.layers.8.input_layernorm.weight

onnx/llama_model.layers.8.post_attention_layernorm.weight

onnx/llama_model.layers.9.input_layernorm.weight

onnx/llama_model.layers.9.post_attention_layernorm.weight

onnx/llama_model.norm.weight

onnx/llm_backbone_v3.onnx.data

onnx/llm_backbone_v4.onnx.data

onnx/llm_backbone_v9.onnx ONNX

onnx/onnx__MatMul_6490

onnx/onnx__MatMul_6491

onnx/onnx__MatMul_6492

onnx/onnx__MatMul_6545

onnx/onnx__MatMul_6546

onnx/onnx__MatMul_6547

onnx/onnx__MatMul_6548

onnx/onnx__MatMul_6549

onnx/onnx__MatMul_6550

onnx/onnx__MatMul_6551

onnx/onnx__MatMul_6598

onnx/onnx__MatMul_6599

onnx/onnx__MatMul_6600

onnx/onnx__MatMul_6601

onnx/onnx__MatMul_6602

onnx/onnx__MatMul_6603

onnx/onnx__MatMul_6604

onnx/onnx__MatMul_6651

onnx/onnx__MatMul_6652

onnx/onnx__MatMul_6653

onnx/onnx__MatMul_6654

onnx/onnx__MatMul_6655

onnx/onnx__MatMul_6656

onnx/onnx__MatMul_6657

onnx/onnx__MatMul_6704

onnx/onnx__MatMul_6705

onnx/onnx__MatMul_6706

onnx/onnx__MatMul_6707

onnx/onnx__MatMul_6708

onnx/onnx__MatMul_6709

onnx/onnx__MatMul_6710

onnx/onnx__MatMul_6757

onnx/onnx__MatMul_6758

onnx/onnx__MatMul_6759

onnx/onnx__MatMul_6760

onnx/onnx__MatMul_6761

onnx/onnx__MatMul_6762

onnx/onnx__MatMul_6763

onnx/onnx__MatMul_6810

onnx/onnx__MatMul_6811

onnx/onnx__MatMul_6812

onnx/onnx__MatMul_6813

onnx/onnx__MatMul_6814

onnx/onnx__MatMul_6815

onnx/onnx__MatMul_6816

onnx/onnx__MatMul_6863

onnx/onnx__MatMul_6864

onnx/onnx__MatMul_6865

onnx/onnx__MatMul_6866

onnx/onnx__MatMul_6867

onnx/onnx__MatMul_6868

onnx/onnx__MatMul_6869

onnx/onnx__MatMul_6916

onnx/onnx__MatMul_6917

onnx/onnx__MatMul_6918

onnx/onnx__MatMul_6919

onnx/onnx__MatMul_6920

onnx/onnx__MatMul_6921

onnx/onnx__MatMul_6922

onnx/onnx__MatMul_6969

onnx/onnx__MatMul_6970

onnx/onnx__MatMul_6971

onnx/onnx__MatMul_6972

onnx/onnx__MatMul_6973

onnx/onnx__MatMul_6974

onnx/onnx__MatMul_6975

onnx/onnx__MatMul_7022

onnx/onnx__MatMul_7023

onnx/onnx__MatMul_7024

onnx/onnx__MatMul_7025

onnx/onnx__MatMul_7026

onnx/onnx__MatMul_7027

onnx/onnx__MatMul_7028

onnx/onnx__MatMul_7075

onnx/onnx__MatMul_7076

onnx/onnx__MatMul_7077

onnx/onnx__MatMul_7078

onnx/onnx__MatMul_7079

onnx/onnx__MatMul_7080

onnx/onnx__MatMul_7081

onnx/onnx__MatMul_7128

onnx/onnx__MatMul_7129

onnx/onnx__MatMul_7130

onnx/onnx__MatMul_7131

onnx/onnx__MatMul_7132

onnx/onnx__MatMul_7133

onnx/onnx__MatMul_7134

onnx/onnx__MatMul_7181

onnx/onnx__MatMul_7182

onnx/onnx__MatMul_7183

onnx/onnx__MatMul_7184

onnx/onnx__MatMul_7185

onnx/onnx__MatMul_7186

onnx/onnx__MatMul_7187

onnx/onnx__MatMul_7234

onnx/onnx__MatMul_7235

onnx/onnx__MatMul_7236

onnx/onnx__MatMul_7237

onnx/onnx__MatMul_7238

onnx/onnx__MatMul_7239

onnx/onnx__MatMul_7240

onnx/onnx__MatMul_7287

onnx/onnx__MatMul_7288

onnx/onnx__MatMul_7289

onnx/onnx__MatMul_7290

onnx/onnx__MatMul_7291

onnx/onnx__MatMul_7292

onnx/onnx__MatMul_7293

onnx/onnx__MatMul_7340

onnx/onnx__MatMul_7341

onnx/onnx__MatMul_7342

onnx/onnx__MatMul_7343

onnx/onnx__MatMul_7344

onnx/onnx__MatMul_7345

onnx/onnx__MatMul_7346

onnx/onnx__MatMul_7393

onnx/onnx__MatMul_7394

onnx/onnx__MatMul_7395

onnx/onnx__MatMul_7396

onnx/onnx__MatMul_7397

onnx/onnx__MatMul_7398

onnx/onnx__MatMul_7399

onnx/onnx__MatMul_7446

onnx/onnx__MatMul_7447

onnx/onnx__MatMul_7448

onnx/onnx__MatMul_7449

onnx/onnx__MatMul_7450

onnx/onnx__MatMul_7451

onnx/onnx__MatMul_7452

onnx/onnx__MatMul_7499

onnx/onnx__MatMul_7500

onnx/onnx__MatMul_7501

onnx/onnx__MatMul_7502

onnx/onnx__MatMul_7503

onnx/onnx__MatMul_7504

onnx/onnx__MatMul_7505

onnx/onnx__MatMul_7552

onnx/onnx__MatMul_7553

onnx/onnx__MatMul_7554

onnx/onnx__MatMul_7555

onnx/onnx__MatMul_7556

onnx/onnx__MatMul_7557

onnx/onnx__MatMul_7558

onnx/onnx__MatMul_7605

onnx/onnx__MatMul_7606

onnx/onnx__MatMul_7607

onnx/onnx__MatMul_7608

onnx/onnx__MatMul_7609

onnx/onnx__MatMul_7610

onnx/onnx__MatMul_7611

onnx/onnx__MatMul_7658

onnx/onnx__MatMul_7659

onnx/onnx__MatMul_7660

onnx/onnx__MatMul_7661

onnx/onnx__MatMul_7662

onnx/onnx__MatMul_7663

onnx/onnx__MatMul_7664

onnx/onnx__MatMul_7711

onnx/onnx__MatMul_7712

onnx/onnx__MatMul_7713

onnx/onnx__MatMul_7714

onnx/onnx__MatMul_7715

onnx/onnx__MatMul_7716

onnx/onnx__MatMul_7717

onnx/onnx__MatMul_7764

onnx/onnx__MatMul_7765

onnx/onnx__MatMul_7766

onnx/onnx__MatMul_7767

onnx/onnx__MatMul_7768

onnx/onnx__MatMul_7769

onnx/onnx__MatMul_7770

onnx/onnx__MatMul_7817

onnx/onnx__MatMul_7818

onnx/onnx__MatMul_7819

onnx/onnx__MatMul_7820

onnx/onnx__MatMul_7821

onnx/onnx__MatMul_7822

onnx/onnx__MatMul_7823

onnx/onnx__MatMul_7870

onnx/onnx__MatMul_7871

onnx/onnx__MatMul_7872

onnx/onnx__MatMul_7873

onnx/onnx__MatMul_7874

onnx/onnx__MatMul_7875

onnx/onnx__MatMul_7876

onnx/onnx__MatMul_7923

onnx/onnx__MatMul_7924

onnx/onnx__MatMul_7925

onnx/onnx__MatMul_7926

onnx/onnx__MatMul_7927

onnx/onnx__MatMul_7928

onnx/onnx__MatMul_7929

onnx/onnx__MatMul_7976

onnx/onnx__MatMul_7977

onnx/onnx__MatMul_7978

onnx/onnx__MatMul_7979

onnx/onnx__MatMul_7980

onnx/onnx__MatMul_7981

onnx/onnx__MatMul_7982

onnx/onnx__MatMul_8029

onnx/onnx__MatMul_8030

onnx/onnx__MatMul_8031

onnx/onnx__MatMul_8032

onnx/onnx__MatMul_8033

onnx/onnx__MatMul_8034

onnx/onnx__MatMul_8035

onnx/onnx__MatMul_8082

onnx/onnx__MatMul_8083

onnx/onnx__MatMul_8084

onnx/onnx__MatMul_8085

onnx/onnx__MatMul_8086

onnx/onnx__MatMul_8087

onnx/onnx__MatMul_8088

onnx/onnx__MatMul_8135

onnx/onnx__MatMul_8136

onnx/onnx__MatMul_8137

onnx/onnx__MatMul_8138

onnx/onnx__MatMul_8139

onnx/onnx__MatMul_8140

onnx/onnx__MatMul_8141

onnx/onnx__MatMul_8188

onnx/onnx__MatMul_8189

onnx/onnx__MatMul_8190

onnx/onnx__MatMul_8191

onnx/projector.fc1.bias

onnx/projector.fc2.bias

onnx/projector.fc3.bias

onnx/vision_backbone.featurizer.blocks.0.attn.proj.bias

onnx/vision_backbone.featurizer.blocks.0.attn.qkv.bias

onnx/vision_backbone.featurizer.blocks.0.ls1.scale_factor

onnx/vision_backbone.featurizer.blocks.0.ls2.scale_factor

onnx/vision_backbone.featurizer.blocks.0.mlp.fc1.bias

onnx/vision_backbone.featurizer.blocks.0.mlp.fc2.bias

onnx/vision_backbone.featurizer.blocks.0.norm1.bias

onnx/vision_backbone.featurizer.blocks.0.norm1.weight

onnx/vision_backbone.featurizer.blocks.0.norm2.bias

onnx/vision_backbone.featurizer.blocks.0.norm2.weight

onnx/vision_backbone.featurizer.blocks.1.attn.proj.bias

onnx/vision_backbone.featurizer.blocks.1.attn.qkv.bias

onnx/vision_backbone.featurizer.blocks.1.ls1.scale_factor

onnx/vision_backbone.featurizer.blocks.1.ls2.scale_factor

onnx/vision_backbone.featurizer.blocks.1.mlp.fc1.bias

onnx/vision_backbone.featurizer.blocks.1.mlp.fc2.bias

onnx/vision_backbone.featurizer.blocks.1.norm1.bias

onnx/vision_backbone.featurizer.blocks.1.norm1.weight

onnx/vision_backbone.featurizer.blocks.1.norm2.bias

onnx/vision_backbone.featurizer.blocks.1.norm2.weight

onnx/vision_backbone.featurizer.blocks.10.attn.proj.bias

onnx/vision_backbone.featurizer.blocks.10.attn.qkv.bias

onnx/vision_backbone.featurizer.blocks.10.ls1.scale_factor

onnx/vision_backbone.featurizer.blocks.10.ls2.scale_factor

onnx/vision_backbone.featurizer.blocks.10.mlp.fc1.bias

onnx/vision_backbone.featurizer.blocks.10.mlp.fc2.bias

onnx/vision_backbone.featurizer.blocks.10.norm1.bias

onnx/vision_backbone.featurizer.blocks.10.norm1.weight

onnx/vision_backbone.featurizer.blocks.10.norm2.bias

onnx/vision_backbone.featurizer.blocks.10.norm2.weight

onnx/vision_backbone.featurizer.blocks.11.attn.proj.bias

onnx/vision_backbone.featurizer.blocks.11.attn.qkv.bias

onnx/vision_backbone.featurizer.blocks.11.ls1.scale_factor

onnx/vision_backbone.featurizer.blocks.11.ls2.scale_factor

onnx/vision_backbone.featurizer.blocks.11.mlp.fc1.bias

onnx/vision_backbone.featurizer.blocks.11.mlp.fc2.bias

onnx/vision_backbone.featurizer.blocks.11.norm1.bias

onnx/vision_backbone.featurizer.blocks.11.norm1.weight

onnx/vision_backbone.featurizer.blocks.11.norm2.bias

onnx/vision_backbone.featurizer.blocks.11.norm2.weight

onnx/vision_backbone.featurizer.blocks.12.attn.proj.bias

onnx/vision_backbone.featurizer.blocks.12.attn.qkv.bias

onnx/vision_backbone.featurizer.blocks.12.ls1.scale_factor

onnx/vision_backbone.featurizer.blocks.12.ls2.scale_factor

onnx/vision_backbone.featurizer.blocks.12.mlp.fc1.bias

onnx/vision_backbone.featurizer.blocks.12.mlp.fc2.bias

onnx/vision_backbone.featurizer.blocks.12.norm1.bias

onnx/vision_backbone.featurizer.blocks.12.norm1.weight

onnx/vision_backbone.featurizer.blocks.12.norm2.bias

onnx/vision_backbone.featurizer.blocks.12.norm2.weight

onnx/vision_backbone.featurizer.blocks.13.attn.proj.bias

onnx/vision_backbone.featurizer.blocks.13.attn.qkv.bias

onnx/vision_backbone.featurizer.blocks.13.ls1.scale_factor

onnx/vision_backbone.featurizer.blocks.13.ls2.scale_factor

onnx/vision_backbone.featurizer.blocks.13.mlp.fc1.bias

onnx/vision_backbone.featurizer.blocks.13.mlp.fc2.bias

onnx/vision_backbone.featurizer.blocks.13.norm1.bias

onnx/vision_backbone.featurizer.blocks.13.norm1.weight

onnx/vision_backbone.featurizer.blocks.13.norm2.bias

onnx/vision_backbone.featurizer.blocks.13.norm2.weight

onnx/vision_backbone.featurizer.blocks.14.attn.proj.bias

onnx/vision_backbone.featurizer.blocks.14.attn.qkv.bias

onnx/vision_backbone.featurizer.blocks.14.ls1.scale_factor

onnx/vision_backbone.featurizer.blocks.14.ls2.scale_factor

onnx/vision_backbone.featurizer.blocks.14.mlp.fc1.bias

onnx/vision_backbone.featurizer.blocks.14.mlp.fc2.bias

onnx/vision_backbone.featurizer.blocks.14.norm1.bias

onnx/vision_backbone.featurizer.blocks.14.norm1.weight

onnx/vision_backbone.featurizer.blocks.14.norm2.bias

onnx/vision_backbone.featurizer.blocks.14.norm2.weight

onnx/vision_backbone.featurizer.blocks.15.attn.proj.bias

onnx/vision_backbone.featurizer.blocks.15.attn.qkv.bias

onnx/vision_backbone.featurizer.blocks.15.ls1.scale_factor

onnx/vision_backbone.featurizer.blocks.15.ls2.scale_factor

onnx/vision_backbone.featurizer.blocks.15.mlp.fc1.bias

onnx/vision_backbone.featurizer.blocks.15.mlp.fc2.bias

onnx/vision_backbone.featurizer.blocks.15.norm1.bias

onnx/vision_backbone.featurizer.blocks.15.norm1.weight

onnx/vision_backbone.featurizer.blocks.15.norm2.bias

onnx/vision_backbone.featurizer.blocks.15.norm2.weight

onnx/vision_backbone.featurizer.blocks.16.attn.proj.bias

onnx/vision_backbone.featurizer.blocks.16.attn.qkv.bias

onnx/vision_backbone.featurizer.blocks.16.ls1.scale_factor

onnx/vision_backbone.featurizer.blocks.16.ls2.scale_factor

onnx/vision_backbone.featurizer.blocks.16.mlp.fc1.bias

onnx/vision_backbone.featurizer.blocks.16.mlp.fc2.bias

onnx/vision_backbone.featurizer.blocks.16.norm1.bias

onnx/vision_backbone.featurizer.blocks.16.norm1.weight

onnx/vision_backbone.featurizer.blocks.16.norm2.bias

onnx/vision_backbone.featurizer.blocks.16.norm2.weight

onnx/vision_backbone.featurizer.blocks.17.attn.proj.bias

onnx/vision_backbone.featurizer.blocks.17.attn.qkv.bias

onnx/vision_backbone.featurizer.blocks.17.ls1.scale_factor

onnx/vision_backbone.featurizer.blocks.17.ls2.scale_factor

onnx/vision_backbone.featurizer.blocks.17.mlp.fc1.bias

onnx/vision_backbone.featurizer.blocks.17.mlp.fc2.bias

onnx/vision_backbone.featurizer.blocks.17.norm1.bias

onnx/vision_backbone.featurizer.blocks.17.norm1.weight

onnx/vision_backbone.featurizer.blocks.17.norm2.bias

onnx/vision_backbone.featurizer.blocks.17.norm2.weight

onnx/vision_backbone.featurizer.blocks.18.attn.proj.bias

onnx/vision_backbone.featurizer.blocks.18.attn.qkv.bias

onnx/vision_backbone.featurizer.blocks.18.ls1.scale_factor

onnx/vision_backbone.featurizer.blocks.18.ls2.scale_factor

onnx/vision_backbone.featurizer.blocks.18.mlp.fc1.bias

onnx/vision_backbone.featurizer.blocks.18.mlp.fc2.bias

onnx/vision_backbone.featurizer.blocks.18.norm1.bias

onnx/vision_backbone.featurizer.blocks.18.norm1.weight

onnx/vision_backbone.featurizer.blocks.18.norm2.bias

onnx/vision_backbone.featurizer.blocks.18.norm2.weight

onnx/vision_backbone.featurizer.blocks.19.attn.proj.bias

onnx/vision_backbone.featurizer.blocks.19.attn.qkv.bias

onnx/vision_backbone.featurizer.blocks.19.ls1.scale_factor

onnx/vision_backbone.featurizer.blocks.19.ls2.scale_factor

onnx/vision_backbone.featurizer.blocks.19.mlp.fc1.bias

onnx/vision_backbone.featurizer.blocks.19.mlp.fc2.bias

onnx/vision_backbone.featurizer.blocks.19.norm1.bias

onnx/vision_backbone.featurizer.blocks.19.norm1.weight

onnx/vision_backbone.featurizer.blocks.19.norm2.bias

onnx/vision_backbone.featurizer.blocks.19.norm2.weight

onnx/vision_backbone.featurizer.blocks.2.attn.proj.bias

onnx/vision_backbone.featurizer.blocks.2.attn.qkv.bias

onnx/vision_backbone.featurizer.blocks.2.ls1.scale_factor

onnx/vision_backbone.featurizer.blocks.2.ls2.scale_factor

onnx/vision_backbone.featurizer.blocks.2.mlp.fc1.bias

onnx/vision_backbone.featurizer.blocks.2.mlp.fc2.bias

onnx/vision_backbone.featurizer.blocks.2.norm1.bias

onnx/vision_backbone.featurizer.blocks.2.norm1.weight

onnx/vision_backbone.featurizer.blocks.2.norm2.bias

onnx/vision_backbone.featurizer.blocks.2.norm2.weight

onnx/vision_backbone.featurizer.blocks.20.attn.proj.bias

onnx/vision_backbone.featurizer.blocks.20.attn.qkv.bias

onnx/vision_backbone.featurizer.blocks.20.ls1.scale_factor

onnx/vision_backbone.featurizer.blocks.20.ls2.scale_factor

onnx/vision_backbone.featurizer.blocks.20.mlp.fc1.bias

onnx/vision_backbone.featurizer.blocks.20.mlp.fc2.bias

onnx/vision_backbone.featurizer.blocks.20.norm1.bias

onnx/vision_backbone.featurizer.blocks.20.norm1.weight

onnx/vision_backbone.featurizer.blocks.20.norm2.bias

onnx/vision_backbone.featurizer.blocks.20.norm2.weight

onnx/vision_backbone.featurizer.blocks.21.attn.proj.bias

onnx/vision_backbone.featurizer.blocks.21.attn.qkv.bias

onnx/vision_backbone.featurizer.blocks.21.ls1.scale_factor

onnx/vision_backbone.featurizer.blocks.21.ls2.scale_factor

onnx/vision_backbone.featurizer.blocks.21.mlp.fc1.bias

onnx/vision_backbone.featurizer.blocks.21.mlp.fc2.bias

onnx/vision_backbone.featurizer.blocks.21.norm1.bias

onnx/vision_backbone.featurizer.blocks.21.norm1.weight

onnx/vision_backbone.featurizer.blocks.21.norm2.bias

onnx/vision_backbone.featurizer.blocks.21.norm2.weight

onnx/vision_backbone.featurizer.blocks.22.attn.proj.bias

onnx/vision_backbone.featurizer.blocks.22.attn.qkv.bias

onnx/vision_backbone.featurizer.blocks.22.ls1.scale_factor

onnx/vision_backbone.featurizer.blocks.22.ls2.scale_factor

onnx/vision_backbone.featurizer.blocks.22.mlp.fc1.bias

onnx/vision_backbone.featurizer.blocks.22.mlp.fc2.bias

onnx/vision_backbone.featurizer.blocks.22.norm1.bias

onnx/vision_backbone.featurizer.blocks.22.norm1.weight

onnx/vision_backbone.featurizer.blocks.22.norm2.bias

onnx/vision_backbone.featurizer.blocks.22.norm2.weight

onnx/vision_backbone.featurizer.blocks.3.attn.proj.bias

onnx/vision_backbone.featurizer.blocks.3.attn.qkv.bias

onnx/vision_backbone.featurizer.blocks.3.ls1.scale_factor

onnx/vision_backbone.featurizer.blocks.3.ls2.scale_factor

onnx/vision_backbone.featurizer.blocks.3.mlp.fc1.bias

onnx/vision_backbone.featurizer.blocks.3.mlp.fc2.bias

onnx/vision_backbone.featurizer.blocks.3.norm1.bias

onnx/vision_backbone.featurizer.blocks.3.norm1.weight

onnx/vision_backbone.featurizer.blocks.3.norm2.bias

onnx/vision_backbone.featurizer.blocks.3.norm2.weight

onnx/vision_backbone.featurizer.blocks.4.attn.proj.bias

onnx/vision_backbone.featurizer.blocks.4.attn.qkv.bias

onnx/vision_backbone.featurizer.blocks.4.ls1.scale_factor

onnx/vision_backbone.featurizer.blocks.4.ls2.scale_factor

onnx/vision_backbone.featurizer.blocks.4.mlp.fc1.bias

onnx/vision_backbone.featurizer.blocks.4.mlp.fc2.bias

onnx/vision_backbone.featurizer.blocks.4.norm1.bias

onnx/vision_backbone.featurizer.blocks.4.norm1.weight

onnx/vision_backbone.featurizer.blocks.4.norm2.bias

onnx/vision_backbone.featurizer.blocks.4.norm2.weight

onnx/vision_backbone.featurizer.blocks.5.attn.proj.bias

onnx/vision_backbone.featurizer.blocks.5.attn.qkv.bias

onnx/vision_backbone.featurizer.blocks.5.ls1.scale_factor

onnx/vision_backbone.featurizer.blocks.5.ls2.scale_factor

onnx/vision_backbone.featurizer.blocks.5.mlp.fc1.bias

onnx/vision_backbone.featurizer.blocks.5.mlp.fc2.bias

onnx/vision_backbone.featurizer.blocks.5.norm1.bias

onnx/vision_backbone.featurizer.blocks.5.norm1.weight

onnx/vision_backbone.featurizer.blocks.5.norm2.bias

onnx/vision_backbone.featurizer.blocks.5.norm2.weight

onnx/vision_backbone.featurizer.blocks.6.attn.proj.bias

onnx/vision_backbone.featurizer.blocks.6.attn.qkv.bias

onnx/vision_backbone.featurizer.blocks.6.ls1.scale_factor

onnx/vision_backbone.featurizer.blocks.6.ls2.scale_factor

onnx/vision_backbone.featurizer.blocks.6.mlp.fc1.bias

onnx/vision_backbone.featurizer.blocks.6.mlp.fc2.bias

onnx/vision_backbone.featurizer.blocks.6.norm1.bias

onnx/vision_backbone.featurizer.blocks.6.norm1.weight

onnx/vision_backbone.featurizer.blocks.6.norm2.bias

onnx/vision_backbone.featurizer.blocks.6.norm2.weight

onnx/vision_backbone.featurizer.blocks.7.attn.proj.bias

onnx/vision_backbone.featurizer.blocks.7.attn.qkv.bias

onnx/vision_backbone.featurizer.blocks.7.ls1.scale_factor

onnx/vision_backbone.featurizer.blocks.7.ls2.scale_factor

onnx/vision_backbone.featurizer.blocks.7.mlp.fc1.bias

onnx/vision_backbone.featurizer.blocks.7.mlp.fc2.bias

onnx/vision_backbone.featurizer.blocks.7.norm1.bias

onnx/vision_backbone.featurizer.blocks.7.norm1.weight

onnx/vision_backbone.featurizer.blocks.7.norm2.bias

onnx/vision_backbone.featurizer.blocks.7.norm2.weight

onnx/vision_backbone.featurizer.blocks.8.attn.proj.bias

onnx/vision_backbone.featurizer.blocks.8.attn.qkv.bias

onnx/vision_backbone.featurizer.blocks.8.ls1.scale_factor

onnx/vision_backbone.featurizer.blocks.8.ls2.scale_factor

onnx/vision_backbone.featurizer.blocks.8.mlp.fc1.bias

onnx/vision_backbone.featurizer.blocks.8.mlp.fc2.bias

onnx/vision_backbone.featurizer.blocks.8.norm1.bias

onnx/vision_backbone.featurizer.blocks.8.norm1.weight

onnx/vision_backbone.featurizer.blocks.8.norm2.bias

onnx/vision_backbone.featurizer.blocks.8.norm2.weight

onnx/vision_backbone.featurizer.blocks.9.attn.proj.bias

onnx/vision_backbone.featurizer.blocks.9.attn.qkv.bias

onnx/vision_backbone.featurizer.blocks.9.ls1.scale_factor

onnx/vision_backbone.featurizer.blocks.9.ls2.scale_factor

onnx/vision_backbone.featurizer.blocks.9.mlp.fc1.bias

onnx/vision_backbone.featurizer.blocks.9.mlp.fc2.bias

onnx/vision_backbone.featurizer.blocks.9.norm1.bias

onnx/vision_backbone.featurizer.blocks.9.norm1.weight

onnx/vision_backbone.featurizer.blocks.9.norm2.bias

onnx/vision_backbone.featurizer.blocks.9.norm2.weight

onnx/vision_backbone.featurizer.cls_token

onnx/vision_backbone.featurizer.patch_embed.proj.bias

onnx/vision_backbone.featurizer.patch_embed.proj.weight

onnx/vision_backbone.featurizer.pos_embed

onnx/vision_backbone.featurizer.reg_token

onnx/vision_backbone.fused_featurizer.blocks.0.attn.proj.bias

onnx/vision_backbone.fused_featurizer.blocks.0.attn.qkv.bias

onnx/vision_backbone.fused_featurizer.blocks.0.mlp.fc1.bias

onnx/vision_backbone.fused_featurizer.blocks.0.mlp.fc2.bias

onnx/vision_backbone.fused_featurizer.blocks.0.norm1.bias

onnx/vision_backbone.fused_featurizer.blocks.0.norm1.weight

onnx/vision_backbone.fused_featurizer.blocks.0.norm2.bias

onnx/vision_backbone.fused_featurizer.blocks.0.norm2.weight

onnx/vision_backbone.fused_featurizer.blocks.1.attn.proj.bias

onnx/vision_backbone.fused_featurizer.blocks.1.attn.qkv.bias

onnx/vision_backbone.fused_featurizer.blocks.1.mlp.fc1.bias

onnx/vision_backbone.fused_featurizer.blocks.1.mlp.fc2.bias

onnx/vision_backbone.fused_featurizer.blocks.1.norm1.bias

onnx/vision_backbone.fused_featurizer.blocks.1.norm1.weight

onnx/vision_backbone.fused_featurizer.blocks.1.norm2.bias

onnx/vision_backbone.fused_featurizer.blocks.1.norm2.weight

onnx/vision_backbone.fused_featurizer.blocks.10.attn.proj.bias

onnx/vision_backbone.fused_featurizer.blocks.10.attn.qkv.bias

onnx/vision_backbone.fused_featurizer.blocks.10.mlp.fc1.bias

onnx/vision_backbone.fused_featurizer.blocks.10.mlp.fc2.bias

onnx/vision_backbone.fused_featurizer.blocks.10.norm1.bias

onnx/vision_backbone.fused_featurizer.blocks.10.norm1.weight

onnx/vision_backbone.fused_featurizer.blocks.10.norm2.bias

onnx/vision_backbone.fused_featurizer.blocks.10.norm2.weight

onnx/vision_backbone.fused_featurizer.blocks.11.attn.proj.bias

onnx/vision_backbone.fused_featurizer.blocks.11.attn.qkv.bias

onnx/vision_backbone.fused_featurizer.blocks.11.mlp.fc1.bias

onnx/vision_backbone.fused_featurizer.blocks.11.mlp.fc2.bias

onnx/vision_backbone.fused_featurizer.blocks.11.norm1.bias

onnx/vision_backbone.fused_featurizer.blocks.11.norm1.weight

onnx/vision_backbone.fused_featurizer.blocks.11.norm2.bias

onnx/vision_backbone.fused_featurizer.blocks.11.norm2.weight

onnx/vision_backbone.fused_featurizer.blocks.12.attn.proj.bias

onnx/vision_backbone.fused_featurizer.blocks.12.attn.qkv.bias

onnx/vision_backbone.fused_featurizer.blocks.12.mlp.fc1.bias

onnx/vision_backbone.fused_featurizer.blocks.12.mlp.fc2.bias

onnx/vision_backbone.fused_featurizer.blocks.12.norm1.bias

onnx/vision_backbone.fused_featurizer.blocks.12.norm1.weight

onnx/vision_backbone.fused_featurizer.blocks.12.norm2.bias

onnx/vision_backbone.fused_featurizer.blocks.12.norm2.weight

onnx/vision_backbone.fused_featurizer.blocks.13.attn.proj.bias

onnx/vision_backbone.fused_featurizer.blocks.13.attn.qkv.bias

onnx/vision_backbone.fused_featurizer.blocks.13.mlp.fc1.bias

onnx/vision_backbone.fused_featurizer.blocks.13.mlp.fc2.bias

onnx/vision_backbone.fused_featurizer.blocks.13.norm1.bias

onnx/vision_backbone.fused_featurizer.blocks.13.norm1.weight

onnx/vision_backbone.fused_featurizer.blocks.13.norm2.bias

onnx/vision_backbone.fused_featurizer.blocks.13.norm2.weight

onnx/vision_backbone.fused_featurizer.blocks.14.attn.proj.bias

onnx/vision_backbone.fused_featurizer.blocks.14.attn.qkv.bias

onnx/vision_backbone.fused_featurizer.blocks.14.mlp.fc1.bias

onnx/vision_backbone.fused_featurizer.blocks.14.mlp.fc2.bias

onnx/vision_backbone.fused_featurizer.blocks.14.norm1.bias

onnx/vision_backbone.fused_featurizer.blocks.14.norm1.weight

onnx/vision_backbone.fused_featurizer.blocks.14.norm2.bias

onnx/vision_backbone.fused_featurizer.blocks.14.norm2.weight

onnx/vision_backbone.fused_featurizer.blocks.15.attn.proj.bias

onnx/vision_backbone.fused_featurizer.blocks.15.attn.qkv.bias

onnx/vision_backbone.fused_featurizer.blocks.15.mlp.fc1.bias

onnx/vision_backbone.fused_featurizer.blocks.15.mlp.fc2.bias

onnx/vision_backbone.fused_featurizer.blocks.15.norm1.bias

onnx/vision_backbone.fused_featurizer.blocks.15.norm1.weight

onnx/vision_backbone.fused_featurizer.blocks.15.norm2.bias

onnx/vision_backbone.fused_featurizer.blocks.15.norm2.weight

onnx/vision_backbone.fused_featurizer.blocks.16.attn.proj.bias

onnx/vision_backbone.fused_featurizer.blocks.16.attn.qkv.bias

onnx/vision_backbone.fused_featurizer.blocks.16.mlp.fc1.bias

onnx/vision_backbone.fused_featurizer.blocks.16.mlp.fc2.bias

onnx/vision_backbone.fused_featurizer.blocks.16.norm1.bias

onnx/vision_backbone.fused_featurizer.blocks.16.norm1.weight

onnx/vision_backbone.fused_featurizer.blocks.16.norm2.bias

onnx/vision_backbone.fused_featurizer.blocks.16.norm2.weight

onnx/vision_backbone.fused_featurizer.blocks.17.attn.proj.bias

onnx/vision_backbone.fused_featurizer.blocks.17.attn.qkv.bias

onnx/vision_backbone.fused_featurizer.blocks.17.mlp.fc1.bias

onnx/vision_backbone.fused_featurizer.blocks.17.mlp.fc2.bias

onnx/vision_backbone.fused_featurizer.blocks.17.norm1.bias

onnx/vision_backbone.fused_featurizer.blocks.17.norm1.weight

onnx/vision_backbone.fused_featurizer.blocks.17.norm2.bias

onnx/vision_backbone.fused_featurizer.blocks.17.norm2.weight

onnx/vision_backbone.fused_featurizer.blocks.18.attn.proj.bias

onnx/vision_backbone.fused_featurizer.blocks.18.attn.qkv.bias

onnx/vision_backbone.fused_featurizer.blocks.18.mlp.fc1.bias

onnx/vision_backbone.fused_featurizer.blocks.18.mlp.fc2.bias

onnx/vision_backbone.fused_featurizer.blocks.18.norm1.bias

onnx/vision_backbone.fused_featurizer.blocks.18.norm1.weight

onnx/vision_backbone.fused_featurizer.blocks.18.norm2.bias

onnx/vision_backbone.fused_featurizer.blocks.18.norm2.weight

onnx/vision_backbone.fused_featurizer.blocks.19.attn.proj.bias

onnx/vision_backbone.fused_featurizer.blocks.19.attn.qkv.bias

onnx/vision_backbone.fused_featurizer.blocks.19.mlp.fc1.bias

onnx/vision_backbone.fused_featurizer.blocks.19.mlp.fc2.bias

onnx/vision_backbone.fused_featurizer.blocks.19.norm1.bias

onnx/vision_backbone.fused_featurizer.blocks.19.norm1.weight

onnx/vision_backbone.fused_featurizer.blocks.19.norm2.bias

onnx/vision_backbone.fused_featurizer.blocks.19.norm2.weight

onnx/vision_backbone.fused_featurizer.blocks.2.attn.proj.bias

onnx/vision_backbone.fused_featurizer.blocks.2.attn.qkv.bias

onnx/vision_backbone.fused_featurizer.blocks.2.mlp.fc1.bias

onnx/vision_backbone.fused_featurizer.blocks.2.mlp.fc2.bias

onnx/vision_backbone.fused_featurizer.blocks.2.norm1.bias

onnx/vision_backbone.fused_featurizer.blocks.2.norm1.weight

onnx/vision_backbone.fused_featurizer.blocks.2.norm2.bias

onnx/vision_backbone.fused_featurizer.blocks.2.norm2.weight

onnx/vision_backbone.fused_featurizer.blocks.20.attn.proj.bias

onnx/vision_backbone.fused_featurizer.blocks.20.attn.qkv.bias

onnx/vision_backbone.fused_featurizer.blocks.20.mlp.fc1.bias

onnx/vision_backbone.fused_featurizer.blocks.20.mlp.fc2.bias

onnx/vision_backbone.fused_featurizer.blocks.20.norm1.bias

onnx/vision_backbone.fused_featurizer.blocks.20.norm1.weight

onnx/vision_backbone.fused_featurizer.blocks.20.norm2.bias

onnx/vision_backbone.fused_featurizer.blocks.20.norm2.weight

onnx/vision_backbone.fused_featurizer.blocks.21.attn.proj.bias

onnx/vision_backbone.fused_featurizer.blocks.21.attn.qkv.bias

onnx/vision_backbone.fused_featurizer.blocks.21.mlp.fc1.bias

onnx/vision_backbone.fused_featurizer.blocks.21.mlp.fc2.bias

onnx/vision_backbone.fused_featurizer.blocks.21.norm1.bias

onnx/vision_backbone.fused_featurizer.blocks.21.norm1.weight

onnx/vision_backbone.fused_featurizer.blocks.21.norm2.bias

onnx/vision_backbone.fused_featurizer.blocks.21.norm2.weight

onnx/vision_backbone.fused_featurizer.blocks.22.attn.proj.bias

onnx/vision_backbone.fused_featurizer.blocks.22.attn.qkv.bias

onnx/vision_backbone.fused_featurizer.blocks.22.mlp.fc1.bias

onnx/vision_backbone.fused_featurizer.blocks.22.mlp.fc2.bias

onnx/vision_backbone.fused_featurizer.blocks.22.norm1.bias

onnx/vision_backbone.fused_featurizer.blocks.22.norm1.weight

onnx/vision_backbone.fused_featurizer.blocks.22.norm2.bias

onnx/vision_backbone.fused_featurizer.blocks.22.norm2.weight

onnx/vision_backbone.fused_featurizer.blocks.23.attn.proj.bias

onnx/vision_backbone.fused_featurizer.blocks.23.attn.qkv.bias

onnx/vision_backbone.fused_featurizer.blocks.23.mlp.fc1.bias

onnx/vision_backbone.fused_featurizer.blocks.23.mlp.fc2.bias

onnx/vision_backbone.fused_featurizer.blocks.23.norm1.bias

onnx/vision_backbone.fused_featurizer.blocks.23.norm1.weight

onnx/vision_backbone.fused_featurizer.blocks.23.norm2.bias

onnx/vision_backbone.fused_featurizer.blocks.23.norm2.weight

onnx/vision_backbone.fused_featurizer.blocks.24.attn.proj.bias

onnx/vision_backbone.fused_featurizer.blocks.24.attn.qkv.bias

onnx/vision_backbone.fused_featurizer.blocks.24.mlp.fc1.bias

onnx/vision_backbone.fused_featurizer.blocks.24.mlp.fc2.bias

onnx/vision_backbone.fused_featurizer.blocks.24.norm1.bias

onnx/vision_backbone.fused_featurizer.blocks.24.norm1.weight

onnx/vision_backbone.fused_featurizer.blocks.24.norm2.bias

onnx/vision_backbone.fused_featurizer.blocks.24.norm2.weight

onnx/vision_backbone.fused_featurizer.blocks.25.attn.proj.bias

onnx/vision_backbone.fused_featurizer.blocks.25.attn.qkv.bias

onnx/vision_backbone.fused_featurizer.blocks.25.mlp.fc1.bias

onnx/vision_backbone.fused_featurizer.blocks.25.mlp.fc2.bias

onnx/vision_backbone.fused_featurizer.blocks.25.norm1.bias

onnx/vision_backbone.fused_featurizer.blocks.25.norm1.weight

onnx/vision_backbone.fused_featurizer.blocks.25.norm2.bias

onnx/vision_backbone.fused_featurizer.blocks.25.norm2.weight

onnx/vision_backbone.fused_featurizer.blocks.3.attn.proj.bias

onnx/vision_backbone.fused_featurizer.blocks.3.attn.qkv.bias

onnx/vision_backbone.fused_featurizer.blocks.3.mlp.fc1.bias

onnx/vision_backbone.fused_featurizer.blocks.3.mlp.fc2.bias

onnx/vision_backbone.fused_featurizer.blocks.3.norm1.bias

onnx/vision_backbone.fused_featurizer.blocks.3.norm1.weight

onnx/vision_backbone.fused_featurizer.blocks.3.norm2.bias

onnx/vision_backbone.fused_featurizer.blocks.3.norm2.weight

onnx/vision_backbone.fused_featurizer.blocks.4.attn.proj.bias

onnx/vision_backbone.fused_featurizer.blocks.4.attn.qkv.bias

onnx/vision_backbone.fused_featurizer.blocks.4.mlp.fc1.bias

onnx/vision_backbone.fused_featurizer.blocks.4.mlp.fc2.bias

onnx/vision_backbone.fused_featurizer.blocks.4.norm1.bias

onnx/vision_backbone.fused_featurizer.blocks.4.norm1.weight

onnx/vision_backbone.fused_featurizer.blocks.4.norm2.bias

onnx/vision_backbone.fused_featurizer.blocks.4.norm2.weight

onnx/vision_backbone.fused_featurizer.blocks.5.attn.proj.bias

onnx/vision_backbone.fused_featurizer.blocks.5.attn.qkv.bias

onnx/vision_backbone.fused_featurizer.blocks.5.mlp.fc1.bias

onnx/vision_backbone.fused_featurizer.blocks.5.mlp.fc2.bias

onnx/vision_backbone.fused_featurizer.blocks.5.norm1.bias

onnx/vision_backbone.fused_featurizer.blocks.5.norm1.weight

onnx/vision_backbone.fused_featurizer.blocks.5.norm2.bias

onnx/vision_backbone.fused_featurizer.blocks.5.norm2.weight

onnx/vision_backbone.fused_featurizer.blocks.6.attn.proj.bias

onnx/vision_backbone.fused_featurizer.blocks.6.attn.qkv.bias

onnx/vision_backbone.fused_featurizer.blocks.6.mlp.fc1.bias

onnx/vision_backbone.fused_featurizer.blocks.6.mlp.fc2.bias

onnx/vision_backbone.fused_featurizer.blocks.6.norm1.bias

onnx/vision_backbone.fused_featurizer.blocks.6.norm1.weight

onnx/vision_backbone.fused_featurizer.blocks.6.norm2.bias

onnx/vision_backbone.fused_featurizer.blocks.6.norm2.weight

onnx/vision_backbone.fused_featurizer.blocks.7.attn.proj.bias

onnx/vision_backbone.fused_featurizer.blocks.7.attn.qkv.bias

onnx/vision_backbone.fused_featurizer.blocks.7.mlp.fc1.bias

onnx/vision_backbone.fused_featurizer.blocks.7.mlp.fc2.bias

onnx/vision_backbone.fused_featurizer.blocks.7.norm1.bias

onnx/vision_backbone.fused_featurizer.blocks.7.norm1.weight

onnx/vision_backbone.fused_featurizer.blocks.7.norm2.bias

onnx/vision_backbone.fused_featurizer.blocks.7.norm2.weight

onnx/vision_backbone.fused_featurizer.blocks.8.attn.proj.bias

onnx/vision_backbone.fused_featurizer.blocks.8.attn.qkv.bias

onnx/vision_backbone.fused_featurizer.blocks.8.mlp.fc1.bias

onnx/vision_backbone.fused_featurizer.blocks.8.mlp.fc2.bias

onnx/vision_backbone.fused_featurizer.blocks.8.norm1.bias

onnx/vision_backbone.fused_featurizer.blocks.8.norm1.weight

onnx/vision_backbone.fused_featurizer.blocks.8.norm2.bias

onnx/vision_backbone.fused_featurizer.blocks.8.norm2.weight

onnx/vision_backbone.fused_featurizer.blocks.9.attn.proj.bias

onnx/vision_backbone.fused_featurizer.blocks.9.attn.qkv.bias

onnx/vision_backbone.fused_featurizer.blocks.9.mlp.fc1.bias

onnx/vision_backbone.fused_featurizer.blocks.9.mlp.fc2.bias

onnx/vision_backbone.fused_featurizer.blocks.9.norm1.bias

onnx/vision_backbone.fused_featurizer.blocks.9.norm1.weight

onnx/vision_backbone.fused_featurizer.blocks.9.norm2.bias

onnx/vision_backbone.fused_featurizer.blocks.9.norm2.weight

onnx/vision_backbone.fused_featurizer.patch_embed.proj.bias

onnx/vision_backbone.fused_featurizer.patch_embed.proj.weight

onnx/vision_backbone.fused_featurizer.pos_embed

onnx/vision_encoder.onnx ONNX

qnn_models/libaction_head_htp.so

qnn_models/libllm_backbone_htp.so

qnn_models/libvision_encoder_sliced_htp.so

Related models

More models

No task type declared.

xpuenabler/OpenVLA-OFT-LIBERO-130-QNN-HTP

What's in this repo

ONNX files

How to use

README

OpenVLA-OFT LIBERO — QNN HTP (FP16, aarch64)

Repository Structure

Prerequisites

Hardware / OS

Software

Python packages (`qnn` conda env)

Required directories

Environment Setup

1. Set up QNN Python environment

2. Verify toolchain

3. Prepare LLM ONNX (v9 patch)

Conversion

Vision Encoder

Action Head

LLM Backbone

Step-by-Step Conversion Detail

Stage 1 — ONNX → QNN source (qnn-onnx-converter)

Stage 2 — Build directory setup

Stage 3 — Compile to aarch64 shared library

Technical Notes

Why FP16 for HTP?

`qnn-model-lib-generator` cannot be used for large models

C++ standard: gnu++20 → c++11

aarch64 large model relocation issue (LLM only)

PYTHONNOUSERSITE=1 is mandatory

Model Information

Model info

Files (1180)

Related models

More from xpuenabler

More models

Popular models

xpuenabler/OpenVLA-OFT-LIBERO-130-QNN-HTP

What's in this repo

ONNX files

How to use

README

OpenVLA-OFT LIBERO — QNN HTP (FP16, aarch64)

Repository Structure

Prerequisites

Hardware / OS

Software

Python packages (qnn conda env)

Required directories

Environment Setup

1. Set up QNN Python environment

2. Verify toolchain

3. Prepare LLM ONNX (v9 patch)

Conversion

Vision Encoder

Action Head

LLM Backbone

Step-by-Step Conversion Detail

Stage 1 — ONNX → QNN source (qnn-onnx-converter)

Stage 2 — Build directory setup

Stage 3 — Compile to aarch64 shared library

Technical Notes

Why FP16 for HTP?

qnn-model-lib-generator cannot be used for large models

C++ standard: gnu++20 → c++11

aarch64 large model relocation issue (LLM only)

PYTHONNOUSERSITE=1 is mandatory

Model Information

Model info

Files (1180)

Related models

More from xpuenabler

More models

Popular models

Python packages (`qnn` conda env)

`qnn-model-lib-generator` cannot be used for large models