Triangle104/Hercules-6.1-Llama-3.1-8B-Q4_K_S-GGUF

This model was converted to GGUF format from Locutusque/Hercules-6.1-Llama-3.1-8B using llama.cpp via the ggml.ai's GGUF-my-repo space. Refer to the original model card for more details on the model.

---

Model details:

Hercules-6.1-Llama-3.1-8B is a fine-tuned language model derived from Llama-3.1-8B. It is specifically designed to excel in instruction following, function calls, and conversational interactions across various scientific and technical domains. This fine-tuning has hercules-v6.1 with enhanced abilities in:

Complex Instruction Following: Understanding and accurately executing multi-step instructions, even those involving specialized terminology.
Function Calling: Seamlessly interpreting and executing function calls, providing appropriate input and output values.
Domain-Specific Knowledge: Engaging in informative and educational conversations about Biology, Chemistry, Physics, Mathematics, Medicine, Computer Science, and more.

Intended Uses & Potential Bias

Hercules-6.1-Llama-3.1-8B is well-suited to the following applications:

Specialized Chatbots: Creating knowledgeable chatbots and conversational agents in scientific and technical fields.
Instructional Assistants: Supporting users with educational and step-by-step guidance in various disciplines.
Code Generation and Execution: Facilitating code execution through function calls, aiding in software development and prototyping.

Important Note: Although Hercules-v6.1 is carefully constructed, it's important to be aware that the underlying data sources may contain biases or reflect harmful stereotypes. Use this model with caution and consider additional measures to mitigate potential biases in its responses. Limitations and Risks

Toxicity: The dataset contains toxic or harmful examples.
Hallucinations and Factual Errors: Like other language models, Llama-3-Hercules-6.0-8B may generate incorrect or misleading information, especially in specialized domains where it lacks sufficient expertise.
Potential for Misuse: The ability to engage in technical conversations and execute function calls could be misused for malicious purposes.

Evaluations Tasks Version Filter n-shot Metric Value Stderr agieval_nous 0.0 none acc ↑ 0.4427 ± 0.0094

• agieval_aqua_rat 1.0 none 0 acc ↑ 0.2913 ± 0.0286 none 0 acc_norm ↑ 0.2480 ± 0.0272
• agieval_logiqa_en 1.0 none 0 acc ↑ 0.3825 ± 0.0191 none 0 acc_norm ↑ 0.3794 ± 0.0190
• agieval_lsat_ar 1.0 none 0 acc ↑ 0.2087 ± 0.0269 none 0 acc_norm ↑ 0.2043 ± 0.0266
• agieval_lsat_lr 1.0 none 0 acc ↑ 0.4431 ± 0.0220 none 0 acc_norm ↑ 0.4000 ± 0.0217
• agieval_lsat_rc 1.0 none 0 acc ↑ 0.6097 ± 0.0298 none 0 acc_norm ↑ 0.5428 ± 0.0304
• agieval_sat_en 1.0 none 0 acc ↑ 0.7621 ± 0.0297 none 0 acc_norm ↑ 0.6942 ± 0.0322
• agieval_sat_en_without_passage 1.0 none 0 acc ↑ 0.4126 ± 0.0344 none 0 acc_norm ↑ 0.3641 ± 0.0336
• agieval_sat_math 1.0 none 0 acc ↑ 0.4318 ± 0.0335 none 0 acc_norm ↑ 0.3500 ± 0.0322

arc_challenge 1.0 none 0 acc ↑ 0.5247 ± 0.0146 none 0 acc_norm ↑ 0.5606 ± 0.0145 eq_bench 2.1 none 0 eqbench ↑ 63.2023 ± 2.6818 none 0 percent_parseable ↑ 98.8304 ± 0.8246 gsm8k 3.0 flexible-extract 5 exact_match ↑ 0.7801 ± 0.0114 strict-match 5 exact_match ↑ 0.7809 ± 0.0114 truthfulqa_mc2 2.0 none 0 acc ↑ 0.5389 ± 0.0150 Open LLM Leaderboard Evaluation Results

Detailed results can be found here Metric Value Avg. 22.40 IFEval (0-Shot) 60.07 BBH (3-Shot) 24.15 MATH Lvl 5 (4-Shot) 15.63 GPQA (0-shot) 1.45 MuSR (0-shot) 3.42 MMLU-PRO (5-shot) 29.65

---

Use with llama.cpp

Install llama.cpp through brew (works on Mac and Linux)

brew install llama.cpp

Invoke the llama.cpp server or the CLI.

CLI:

llama-cli --hf-repo Triangle104/Hercules-6.1-Llama-3.1-8B-Q4_K_S-GGUF --hf-file hercules-6.1-llama-3.1-8b-q4_k_s.gguf -p "The meaning to life and the universe is"

Server:

llama-server --hf-repo Triangle104/Hercules-6.1-Llama-3.1-8B-Q4_K_S-GGUF --hf-file hercules-6.1-llama-3.1-8b-q4_k_s.gguf -c 2048

Note: You can also use this checkpoint directly through the usage steps listed in the Llama.cpp repo as well.

Step 1: Clone llama.cpp from GitHub.

git clone https://github.com/ggerganov/llama.cpp

Step 2: Move into the llama.cpp folder and build it with LLAMA_CURL=1 flag along with other hardware-specific flags (for ex: LLAMA_CUDA=1 for Nvidia GPUs on Linux).

cd llama.cpp && LLAMA_CURL=1 make

Step 3: Run inference through the main binary.

./llama-cli --hf-repo Triangle104/Hercules-6.1-Llama-3.1-8B-Q4_K_S-GGUF --hf-file hercules-6.1-llama-3.1-8b-q4_k_s.gguf -p "The meaning to life and the universe is"

or

./llama-server --hf-repo Triangle104/Hercules-6.1-Llama-3.1-8B-Q4_K_S-GGUF --hf-file hercules-6.1-llama-3.1-8b-q4_k_s.gguf -c 2048