mlx-examples/llms/mlx_lm
L 0164d2058b
feat: DeepSeek MoE v1 (#942)
* feat: deepseek v1

DeepSeek is still releasing models on the DeepSeek V1 architecture.

```sh
mlx_lm.convert --hf-path deepseek-ai/DeepSeek-Prover-V1.5-RL --mlx-path DeepSeek-Prover-V1.5-RL-8bit --q-bits 8 -q
mlx_lm.generate --model DeepSeek-Prover-V1.5-RL-8bit --ignore-chat-template --max-tokens 512 --prompt 'import Mathlib
import Aesop

set_option maxHeartbeats 0

open BigOperators Real Nat Topology Rat

/-- The second and fourth terms of a geometric sequence are $2$ and $6$. Which of the following is a possible first term?
Show that it is $\frac{2\sqrt{3}}{3}$.-/
theorem amc12b_2003_p6 (a r : ℝ) (u : ℕ → ℝ) (h₀ : ∀ k, u k = a * r ^ k) (h₁ : u 1 = 2)
  (h₂ : u 3 = 6) : u 0 = 2 / Real.sqrt 3 ∨ u 0 = -(2 / Real.sqrt 3) := by'
```

* nits

* nits

* nits

---------

Co-authored-by: Awni Hannun <awni@apple.com>
2024-08-17 07:18:09 -07:00
..
examples Example of response generation with optional arguments (#853) 2024-07-09 06:49:59 -07:00
models feat: DeepSeek MoE v1 (#942) 2024-08-17 07:18:09 -07:00
tuner feat: DeepSeek MoE v1 (#942) 2024-08-17 07:18:09 -07:00
__init__.py mlx_lm: Add Streaming Capability to Generate Function (#807) 2024-06-03 09:04:39 -07:00
convert.py Create executables for generate, lora, server, merge, convert (#682) 2024-04-16 16:08:49 -07:00
fuse.py Allow the entire model to be targed for LoRA and DoRA fine tuning: LoRA and DoRA embeddings with small DoRALinear bug fix (#914) 2024-08-16 07:38:36 -07:00
generate.py Handle longer prompt/generation (#931) 2024-08-16 15:28:39 -07:00
gguf.py Whisper updates to allow HF models (#923) 2024-08-09 11:11:58 -07:00
LORA.md Configuration-based use of HF hub-hosted datasets for training (#701) 2024-06-26 10:20:50 -07:00
lora.py Pass use_dora parameter to linear_to_lora_layers (#885) 2024-07-11 14:34:34 -07:00
MANAGE.md Add model management functionality for local caches (#736) 2024-05-03 12:20:13 -07:00
manage.py Add model management functionality for local caches (#736) 2024-05-03 12:20:13 -07:00
MERGE.md Create executables for generate, lora, server, merge, convert (#682) 2024-04-16 16:08:49 -07:00
merge.py Create executables for generate, lora, server, merge, convert (#682) 2024-04-16 16:08:49 -07:00
py.typed Add py.typed to support PEP-561 (type-hinting) (#389) 2024-01-30 21:17:38 -08:00
README.md feat: move lora into mlx-lm (#337) 2024-01-23 08:44:37 -08:00
requirements.txt Example of response generation with optional arguments (#853) 2024-07-09 06:49:59 -07:00
sample_utils.py Min P implementation (#926) 2024-08-15 15:45:02 -07:00
SERVER.md Adapters loading (#902) 2024-08-01 16:18:18 -07:00
server.py Predict stop sequence matches during streaming (#541) 2024-08-06 15:24:15 -07:00
tokenizer_utils.py fix yi (#852) 2024-06-27 06:38:19 -07:00
UPLOAD.md Mlx llm package (#301) 2024-01-12 10:25:56 -08:00
utils.py Handle longer prompt/generation (#931) 2024-08-16 15:28:39 -07:00
version.py Handle longer prompt/generation (#931) 2024-08-16 15:28:39 -07:00

Generate Text with MLX and 🤗 Hugging Face

This an example of large language model text generation that can pull models from the Hugging Face Hub.

For more information on this example, see the README in the parent directory.

This package also supports fine tuning with LoRA or QLoRA. For more information see the LoRA documentation.