Soraya Hoggarth PRO

somersetent

·

AI & ML interests

None yet

Recent Activity

updated a collection 3 days ago

2-bit Quantization

updated a collection 3 days ago

3-bit Quantization

updated a collection 3 days ago

4-bit Quantization

View all activity

Organizations

None yet

updated 11 collections 3 days ago

2-bit Quantization

2-bit quantization reduces the memory required to store each model weight from 16 bits down to 2 bits. This compresses the model size by roughly 8x. • 37 items • Updated 3 days ago

3-bit Quantization

3-bit quantization in LLMs means compressing model weights from 16-bit precision down to an average of 3 bits per weight. • 37 items • Updated 3 days ago

4-bit Quantization

4-bit in LLMs means compressing model weights from 16-bit to 4-bit precision. This cuts the model's memory size by about 75%. • 63 items • Updated 3 days ago

5-bit Quantization

5-bit quantization in LLMs means compressing the model's weights from high precision (like 16-bit) down to an average of 5 bits per weight. • 36 items • Updated 3 days ago

6-bit Quantization

6-bit quantization in LLMs is a compression method that shrinks model size and memory use by reducing the precision of its weights. • 37 items • Updated 3 days ago

8-bit Quantization

In LLMs, 8-bit means storing the model's weight numbers using 8 bits of computer memory instead of the standard 16 bits. • 49 items • Updated 3 days ago

Instruct

Instructing LLMs means fine-tuning or prompting a base model with explicit directions or tasks so that it acts as a helpful assistant. • 263 items • Updated 3 days ago

Importance Matrix (Imatrix)

The Importance Matrix (iMatrix) is a data-driven calibration method used in low-bit quantization for LLMs. • 334 items • Updated 3 days ago

i1 (llama.cpp and HuggingFace Community)

In the context of local LLMs, quantization and file naming conventions, i1 designates a model quantized using an importance matrix (imatrix). • 121 items • Updated 3 days ago

Meta-Llama

429 items • Updated 3 days ago • 1

GPT-Generated Unified Format (GGUF)

GPT-Generated Unified Format (GGUF) is a single-file binary format used to store and run large language models efficiently on consumer hardware. • 718 items • Updated 3 days ago

liked a model 3 days ago

mradermacher/L3.1-Instruct-Guru-8B-i1-GGUF

8B • Updated May 29, 2025 • 1.59k • 3

updated 8 collections 3 days ago

GPT-Generated Unified Format (GGUF)

GPT-Generated Unified Format (GGUF) is a single-file binary format used to store and run large language models efficiently on consumer hardware. • 718 items • Updated 3 days ago

Instruct

Instructing LLMs means fine-tuning or prompting a base model with explicit directions or tasks so that it acts as a helpful assistant. • 263 items • Updated 3 days ago

Meta-Llama

429 items • Updated 3 days ago • 1

2-bit Quantization

2-bit quantization reduces the memory required to store each model weight from 16 bits down to 2 bits. This compresses the model size by roughly 8x. • 37 items • Updated 3 days ago

3-bit Quantization

3-bit quantization in LLMs means compressing model weights from 16-bit precision down to an average of 3 bits per weight. • 37 items • Updated 3 days ago

4-bit Quantization

4-bit in LLMs means compressing model weights from 16-bit to 4-bit precision. This cuts the model's memory size by about 75%. • 63 items • Updated 3 days ago

5-bit Quantization

5-bit quantization in LLMs means compressing the model's weights from high precision (like 16-bit) down to an average of 5 bits per weight. • 36 items • Updated 3 days ago

6-bit Quantization

6-bit quantization in LLMs is a compression method that shrinks model size and memory use by reducing the precision of its weights. • 37 items • Updated 3 days ago