Index-Translate-2B โ€” MLX 6-bit

English ยท ไธญๆ–‡

English

IndexTeam/Index-Translate-2B converted to MLX for Apple silicon:

  • Text only. The official model also reads images. This repo keeps only the language model; the vision encoder is left out.
  • 6-bit weights, group size 32. The weights are quantized to 6 bits in groups of 32 (7.0 bits per weight with the scales). Other parameters stay bfloat16.
  • Same tokenizer and chat template as the official model.

Files

File What Size
model.safetensors 6-bit weights 1,647 MB
model.safetensors.index.json Weight index
config.json Model configuration, with the quantization settings
generation_config.json End-of-text token ids
tokenizer.json The official tokenizer, saved again by the converter 20 MB
tokenizer_config.json The official file, unchanged
chat_template.jinja The official chat template
LICENSE Apache 2.0
SHA256SUMS Checksums of every file

Sizes are in MB = 10โถ bytes.

The converter saves the tokenizer with <think>, </think> and the other special tokens listed in tokenizer.json itself, so loaders that read only tokenizer.json keep them whole. On our 374 test prompts it gives the same token ids as the official tokenizer. tokenizer_config.json is the official file: the converter's copy names a tokenizer class (Qwen3_5Tokenizer) that swift-transformers does not know.

Use

Index-Translate takes one user message and no system message. Use greedy decoding (temperature 0) and turn thinking off, as the official model card does. The prompt names the languages in Chinese:

่ฏทๅฐ†ไปฅไธ‹{source}ๆ–‡ๆœฌ็ฟป่ฏ‘ไธบ{target}๏ผŒ็›ดๆŽฅ่พ“ๅ‡บ็ฟป่ฏ‘็ป“ๆžœ๏ผŒไธ่ฆ่ฟ›่กŒไปปไฝ•่งฃ้‡Šใ€‚

{text}

For example {source} = ่‹ฑ่ฏญ (English) and {target} = ไธญๆ–‡ (Chinese). Other names: ๆ—ฅ่ฏญ (Japanese), ๅพท่ฏญ (German), ๆณ•่ฏญ (French), ่ฅฟ็ญ็‰™่ฏญ (Spanish), ้Ÿฉ่ฏญ (Korean), ไฟ„่ฏญ (Russian). See the official model card for glossaries and other constraints.

With the chat template and thinking off, the text the model sees is:

<|im_start|>user\n{prompt}<|im_end|>\n<|im_start|>assistant\n<think>\n\n</think>\n\n
from mlx_lm import load, generate

model, tokenizer = load("OpenInfinity/Index-Translate-2B-MLX-6bit")
prompt = "่ฏทๅฐ†ไปฅไธ‹่‹ฑ่ฏญๆ–‡ๆœฌ็ฟป่ฏ‘ไธบไธญๆ–‡๏ผŒ็›ดๆŽฅ่พ“ๅ‡บ็ฟป่ฏ‘็ป“ๆžœ๏ผŒไธ่ฆ่ฟ›่กŒไปปไฝ•่งฃ้‡Šใ€‚\n\nThe quick brown fox jumps over the lazy dog."
text = tokenizer.apply_chat_template([{"role": "user", "content": prompt}],
                                     add_generation_prompt=True, enable_thinking=False, tokenize=False)
print(generate(model, tokenizer, text, max_tokens=1024))
import Foundation
import MLXLLM
import MLXLMCommon

// modelDirectory: a local copy of this repo. tokenizerLoader: a TokenizerLoader from one of
// mlx-swift-lm's tokenizer integrations (mlx-swift-lm 3 does not bundle one).
// Tested with mlx-swift-lm 3.32.3 and swift-transformers 1.3.4.
let container = try await LLMModelFactory.shared.loadContainer(from: modelDirectory, using: tokenizerLoader)
let session = ChatSession(container,
                          generateParameters: GenerateParameters(maxTokens: 1024, temperature: 0),
                          additionalContext: ["enable_thinking": false])
let prompt = "่ฏทๅฐ†ไปฅไธ‹่‹ฑ่ฏญๆ–‡ๆœฌ็ฟป่ฏ‘ไธบไธญๆ–‡๏ผŒ็›ดๆŽฅ่พ“ๅ‡บ็ฟป่ฏ‘็ป“ๆžœ๏ผŒไธ่ฆ่ฟ›่กŒไปปไฝ•่งฃ้‡Šใ€‚\n\nThe quick brown fox jumps over the lazy dog."
print(try await session.respond(to: prompt))

How it was converted

From the official weights (revision 9bc0844531) with mlx-lm 0.32.2 (MLX 0.32.4):

mlx_lm.convert --hf-path IndexTeam/Index-Translate-2B --mlx-path Index-Translate-2B-MLX-6bit \
  -q --q-bits 6 --q-group-size 32

mlx-lm drops the vision encoder when it converts this model. We chose group size 32 over the default 64: with 64, Chinese output drifted further from the original (see below).

Validation

Test set. 374 text segments taken from 12 real web pages with our browser extension's segmenter. They include links marked as <a1>โ€ฆ</a1> and numbers that must stay unchanged, as the extension sends them.

Direction Segments Pages
English โ†’ Chinese 197 Wikipedia, Hacker News, MDN, Paul Graham, Stack Overflow, GitHub, BBC News, arXiv
Chinese โ†’ English 77 Chinese Wikipedia, a Chinese blog
English โ†’ Japanese 50 Wikipedia
English โ†’ German 50 Wikipedia

Each segment used the same prompt for every model, built by our app's translation code, with greedy decoding and thinking off. The reference is the official weights converted to MLX without quantization (bfloat16). We compare against the official Q6_K GGUF, run with llama.cpp (build 11540).

Closeness to the original (chrF++ against the bfloat16 reference; 100 = identical):

Model Size All en โ†’ zh zh โ†’ en en โ†’ ja en โ†’ de Identical outputs
Official GGUF Q6_K 1,606 MB 94.2 92.1 93.1 94.6 98.2 294 / 374
MLX 6-bit, group size 64 1,530 MB 93.7 89.7 96.3 94.1 97.2 282 / 374
This model (6-bit, group size 32) 1,647 MB 94.7 91.8 96.7 94.3 97.9 285 / 374

Checks. Counts of segments with a problem:

Check bfloat16 GGUF Q6_K This model
Link markers lost or changed (of 140 with links) 3 2 4
A number or string that must stay unchanged was changed (of 117) 13 12 12

The original model misses these too, so most of them come from the model, not from quantization. We also read the 100 segments where this model and the GGUF differ; nearly all differ only in wording, and each side has a few errors the other does not.

Speed (MacBook Pro, M3 Max, 128 GB, macOS 27.2; one segment at a time; this model run with mlx-lm in Python, the GGUF with llama-server):

GGUF Q6_K This model
All 374 segments 114.6 s 94.2 s
Generation speed, median 125 tokens/s 169 tokens/s
Time to first token, median 42 ms 35 ms
Load time 4.8 s 0.9 s
Peak memory (process) 1.8 GB 2.1 GB

Many web segments are short, so the time per request weighs more than generation speed: the whole set is 18% faster.

Limitations

  • Tested on one machine only: M3 Max with macOS 27.2.
  • Text only. For image translation use the official weights.
  • Small test set. 374 segments from 12 pages, scored against the original model rather than against human translations.
  • No Traditional Chinese. Like the official model, which does not list it.

License and attribution

  • License. Apache 2.0, the same as Index-Translate-2B.
  • Index-Translate. The model is by IndexTeam, bilibili.
  • This repo. We did the MLX conversion, the quantization and the tests. This repo is not affiliated with IndexTeam.

ไธญๆ–‡

ๆœฌไป“ๅบ“ๆ˜ฏ IndexTeam/Index-Translate-2B ็š„ MLX ็‰ˆๆœฌ๏ผŒ็”จไบŽ Apple ่Šฏ็‰‡๏ผš

  • ๅชๆœ‰ๆ–‡ๅญ—้ƒจๅˆ†๏ผšๅฎ˜ๆ–นๆจกๅž‹ไนŸ่ƒฝ็œ‹ๅ›พใ€‚ๆœฌไป“ๅบ“ๅชไฟ็•™่ฏญ่จ€ๆจกๅž‹๏ผŒไธๅซ่ง†่ง‰็ผ–็ ๅ™จใ€‚
  • 6 bit ๆƒ้‡๏ผŒๆฏ็ป„ 32 ไธช๏ผšๆƒ้‡ๆŒ‰ๆฏ 32 ไธชไธ€็ป„้‡ๅŒ–ไธบ 6 bit๏ผˆ็ฎ—ไธŠ็ผฉๆ”พ็ณปๆ•ฐ็บฆ 7.0 bit/ๆƒ้‡๏ผ‰๏ผŒๅ…ถไฝ™ๅ‚ๆ•ฐไฟๆŒ bfloat16ใ€‚
  • ๅˆ†่ฏๅ™จๅ’Œๅฏน่ฏๆจกๆฟไธŽๅฎ˜ๆ–นๆจกๅž‹็›ธๅŒใ€‚

ๆ–‡ไปถ

ๆ–‡ไปถ ๅ†…ๅฎน ๅคงๅฐ
model.safetensors 6 bit ๆƒ้‡ 1,647 MB
model.safetensors.index.json ๆƒ้‡็ดขๅผ•
config.json ๆจกๅž‹้…็ฝฎ๏ผŒๅซ้‡ๅŒ–่ฎพ็ฝฎ
generation_config.json ็ป“ๆŸ็ฌฆ id
tokenizer.json ๅฎ˜ๆ–นๅˆ†่ฏๅ™จ๏ผŒ็”ฑ่ฝฌๆขๅทฅๅ…ท้‡ๆ–ฐไฟๅญ˜ 20 MB
tokenizer_config.json ๅฎ˜ๆ–นๅŽŸๆ–‡ไปถ๏ผŒๆœชๆ”นๅŠจ
chat_template.jinja ๅฎ˜ๆ–นๅฏน่ฏๆจกๆฟ
LICENSE Apache 2.0
SHA256SUMS ๆ‰€ๆœ‰ๆ–‡ไปถ็š„ๆ ก้ชŒๅ’Œ

ๆœฌ้กต็š„ MB ๆŒ‡ 10โถ ๅญ—่Š‚ใ€‚

่ฝฌๆขๅทฅๅ…ทไฟๅญ˜ๅˆ†่ฏๅ™จๆ—ถ๏ผŒๆŠŠ <think>ใ€</think> ็ญ‰็‰นๆฎŠๆ ‡่ฎฐ็›ดๆŽฅๅ†™่ฟ›ไบ† tokenizer.json๏ผŒๅช่ฏป่ฟ™ไธชๆ–‡ไปถ็š„ๅŠ ่ฝฝๆ–นๅผไนŸไธไผšๆŠŠๅฎƒไปฌๆ‹†ๅผ€ใ€‚ๅœจๆˆ‘ไปฌ็š„ 374 ๆกๆต‹่ฏ•ๆ็คบ่ฏไธŠ๏ผŒๅˆ‡ๅˆ†็ป“ๆžœๅ’Œๅฎ˜ๆ–นๅˆ†่ฏๅ™จๅฎŒๅ…จไธ€่‡ดใ€‚tokenizer_config.json ็”จ็š„ๆ˜ฏๅฎ˜ๆ–นๅŽŸๆ–‡ไปถ๏ผš่ฝฌๆขๅทฅๅ…ทไฟๅญ˜็š„็‰ˆๆœฌๅ†™็š„ๅˆ†่ฏๅ™จ็ฑปๅ๏ผˆQwen3_5Tokenizer๏ผ‰swift-transformers ไธ่ฎค่ฏ†ใ€‚

็”จๆณ•

Index-Translate ๅชๆŽฅๆ”ถไธ€ๆก็”จๆˆทๆถˆๆฏ๏ผŒไธๅŠ ็ณป็ปŸๆถˆๆฏใ€‚ๆŒ‰ๅฎ˜ๆ–นๆจกๅž‹ๅก๏ผŒ็”จ่ดชๅฟƒ่งฃ็ ๏ผˆtemperature 0๏ผ‰ๅนถๅ…ณ้—ญๆ€่€ƒใ€‚ๆ็คบ่ฏ้‡Œ็š„่ฏญ่จ€ๅ็”จไธญๆ–‡๏ผš

่ฏทๅฐ†ไปฅไธ‹{source}ๆ–‡ๆœฌ็ฟป่ฏ‘ไธบ{target}๏ผŒ็›ดๆŽฅ่พ“ๅ‡บ็ฟป่ฏ‘็ป“ๆžœ๏ผŒไธ่ฆ่ฟ›่กŒไปปไฝ•่งฃ้‡Šใ€‚

{text}

ไพ‹ๅฆ‚ {source} = ่‹ฑ่ฏญ๏ผŒ{target} = ไธญๆ–‡ใ€‚ๅ…ถไป–่ฏญ่จ€ๅ๏ผšๆ—ฅ่ฏญใ€ๅพท่ฏญใ€ๆณ•่ฏญใ€่ฅฟ็ญ็‰™่ฏญใ€้Ÿฉ่ฏญใ€ไฟ„่ฏญใ€‚ๆœฏ่ฏญ่กจ็ญ‰็บฆๆŸ็š„ๅ†™ๆณ•่งๅฎ˜ๆ–นๆจกๅž‹ๅกใ€‚

ๅฅ—ไธŠๅฏน่ฏๆจกๆฟใ€ๅ…ณ้—ญๆ€่€ƒๅŽ๏ผŒๆจกๅž‹ๅฎž้™…็œ‹ๅˆฐ็š„ๆ–‡ๆœฌๆ˜ฏ๏ผš

<|im_start|>user\n{prompt}<|im_end|>\n<|im_start|>assistant\n<think>\n\n</think>\n\n
from mlx_lm import load, generate

model, tokenizer = load("OpenInfinity/Index-Translate-2B-MLX-6bit")
prompt = "่ฏทๅฐ†ไปฅไธ‹่‹ฑ่ฏญๆ–‡ๆœฌ็ฟป่ฏ‘ไธบไธญๆ–‡๏ผŒ็›ดๆŽฅ่พ“ๅ‡บ็ฟป่ฏ‘็ป“ๆžœ๏ผŒไธ่ฆ่ฟ›่กŒไปปไฝ•่งฃ้‡Šใ€‚\n\nThe quick brown fox jumps over the lazy dog."
text = tokenizer.apply_chat_template([{"role": "user", "content": prompt}],
                                     add_generation_prompt=True, enable_thinking=False, tokenize=False)
print(generate(model, tokenizer, text, max_tokens=1024))
import Foundation
import MLXLLM
import MLXLMCommon

// modelDirectory๏ผšๆœฌไป“ๅบ“ไธ‹่ฝฝๅˆฐๆœฌๅœฐ็š„็›ฎๅฝ•ใ€‚tokenizerLoader๏ผšmlx-swift-lm ๆŸไธชๅˆ†่ฏๅ™จ้›†ๆˆๅŒ…ๆไพ›็š„
// TokenizerLoader๏ผˆmlx-swift-lm 3 ไธ่‡ชๅธฆๅˆ†่ฏๅ™จ๏ผ‰ใ€‚
// ๅทฒ็”จ mlx-swift-lm 3.32.3 ๅ’Œ swift-transformers 1.3.4 ๅฎžๆต‹ใ€‚
let container = try await LLMModelFactory.shared.loadContainer(from: modelDirectory, using: tokenizerLoader)
let session = ChatSession(container,
                          generateParameters: GenerateParameters(maxTokens: 1024, temperature: 0),
                          additionalContext: ["enable_thinking": false])
let prompt = "่ฏทๅฐ†ไปฅไธ‹่‹ฑ่ฏญๆ–‡ๆœฌ็ฟป่ฏ‘ไธบไธญๆ–‡๏ผŒ็›ดๆŽฅ่พ“ๅ‡บ็ฟป่ฏ‘็ป“ๆžœ๏ผŒไธ่ฆ่ฟ›่กŒไปปไฝ•่งฃ้‡Šใ€‚\n\nThe quick brown fox jumps over the lazy dog."
print(try await session.respond(to: prompt))

่ฝฌๆขๆ–นๆณ•

ไปŽๅฎ˜ๆ–นๆƒ้‡๏ผˆ็‰ˆๆœฌ 9bc0844531๏ผ‰็”จ mlx-lm 0.32.2๏ผˆMLX 0.32.4๏ผ‰่ฝฌๆข๏ผš

mlx_lm.convert --hf-path IndexTeam/Index-Translate-2B --mlx-path Index-Translate-2B-MLX-6bit \
  -q --q-bits 6 --q-group-size 32

mlx-lm ่ฝฌๆข่ฟ™ไธชๆจกๅž‹ๆ—ถไผšๅŽปๆމ่ง†่ง‰็ผ–็ ๅ™จใ€‚ๆˆ‘ไปฌ้€‰ไบ†ๆฏ็ป„ 32 ไธชๆƒ้‡๏ผŒ่€Œไธๆ˜ฏ้ป˜่ฎค็š„ 64 ไธช๏ผš็”จ 64 ๆ—ถ๏ผŒไธญๆ–‡่พ“ๅ‡บๅ’ŒๅŽŸ็‰ˆ็š„ๅทฎ่ทๆ›ดๅคง๏ผˆ่งไธ‹ๆ–‡๏ผ‰ใ€‚

้ชŒ่ฏ

ๆต‹่ฏ•้›†๏ผš็”จๆˆ‘ไปฌๆต่งˆๅ™จๆ‰ฉๅฑ•็š„ๅˆ†ๆฎต้€ป่พ‘๏ผŒไปŽ 12 ไธช็œŸๅฎž็ฝ‘้กตๅ–ไบ† 374 ๆฎตๆ–‡ๅญ—ใ€‚ๅ…ถไธญๅธฆๆœ‰ <a1>โ€ฆ</a1> ๅฝขๅผ็š„้“พๆŽฅๆ ‡่ฎฐ๏ผŒไปฅๅŠๅฟ…้กปๅŽŸๆ ทไฟ็•™็š„ๆ•ฐๅญ—๏ผŒๅ’Œๆ‰ฉๅฑ•ๅฎž้™…ๅ‘ๅ‡บ็š„ไธ€ๆ ทใ€‚

ๆ–นๅ‘ ๆฎตๆ•ฐ ็ฝ‘้กต
่‹ฑ่ฏ‘ไธญ 197 ็ปดๅŸบ็™พ็ง‘ใ€Hacker Newsใ€MDNใ€Paul Grahamใ€Stack Overflowใ€GitHubใ€BBC ๆ–ฐ้—ปใ€arXiv
ไธญ่ฏ‘่‹ฑ 77 ไธญๆ–‡็ปดๅŸบ็™พ็ง‘ใ€ไธ€ไธชไธญๆ–‡ๅšๅฎข
่‹ฑ่ฏ‘ๆ—ฅ 50 ็ปดๅŸบ็™พ็ง‘
่‹ฑ่ฏ‘ๅพท 50 ็ปดๅŸบ็™พ็ง‘

ๆฏๆฎตๅฏนๆ‰€ๆœ‰ๆจกๅž‹็”จๅŒไธ€ๆกๆ็คบ่ฏ๏ผˆ็”ฑๆˆ‘ไปฌๅบ”็”จ็š„็ฟป่ฏ‘ไปฃ็ ็”Ÿๆˆ๏ผ‰๏ผŒ่ดชๅฟƒ่งฃ็ ใ€ๅ…ณ้—ญๆ€่€ƒใ€‚ๅ‚็…งๆ˜ฏๅฎ˜ๆ–นๆƒ้‡ไธ้‡ๅŒ–็›ดๆŽฅ่ฝฌๆˆ MLX๏ผˆbfloat16๏ผ‰ใ€‚ๅฏนๆฏ”ๅฏน่ฑกๆ˜ฏๅฎ˜ๆ–น็š„ Q6_K GGUF๏ผŒ็”จ llama.cpp๏ผˆbuild 11540๏ผ‰่ฟ่กŒใ€‚

ๅ’ŒๅŽŸ็‰ˆ็š„ๆŽฅ่ฟ‘็จ‹ๅบฆ๏ผˆไปฅ bfloat16 ๅ‚็…ง่ฎก็ฎ— chrF++๏ผ›100 ่กจ็คบๅฎŒๅ…จ็›ธๅŒ๏ผ‰๏ผš

ๆจกๅž‹ ๅคงๅฐ ๅ…จ้ƒจ ่‹ฑ่ฏ‘ไธญ ไธญ่ฏ‘่‹ฑ ่‹ฑ่ฏ‘ๆ—ฅ ่‹ฑ่ฏ‘ๅพท ไธŽๅ‚็…งๅฎŒๅ…จ็›ธๅŒ
ๅฎ˜ๆ–น GGUF Q6_K 1,606 MB 94.2 92.1 93.1 94.6 98.2 294 / 374
MLX 6 bit๏ผŒๆฏ็ป„ 64 ไธช 1,530 MB 93.7 89.7 96.3 94.1 97.2 282 / 374
ๆœฌๆจกๅž‹๏ผˆ6 bit๏ผŒๆฏ็ป„ 32 ไธช๏ผ‰ 1,647 MB 94.7 91.8 96.7 94.3 97.9 285 / 374

ๆฃ€ๆŸฅ๏ผšๅ‡บ็Žฐ้—ฎ้ข˜็š„ๆฎตๆ•ฐ๏ผš

ๆฃ€ๆŸฅ้กน bfloat16 GGUF Q6_K ๆœฌๆจกๅž‹
้“พๆŽฅๆ ‡่ฎฐไธขๅคฑๆˆ–ๆ”นๅŠจ๏ผˆๅ…ฑ 140 ๆฎตๅธฆ้“พๆŽฅ๏ผ‰ 3 2 4
ๅฟ…้กปๅŽŸๆ ทไฟ็•™็š„ๆ•ฐๅญ—ๆˆ–ๅญ—็ฌฆไธฒ่ขซๆ”นๅŠจ๏ผˆๅ…ฑ 117 ๆฎต๏ผ‰ 13 12 12

ๅŽŸ็‰ˆๆจกๅž‹ไนŸไผšๅ‡บ่ฟ™ไบ›้—ฎ้ข˜๏ผŒๆ‰€ไปฅๅคง้ƒจๅˆ†ๆฅ่‡ชๆจกๅž‹ๆœฌ่บซ๏ผŒไธๆ˜ฏ้‡ๅŒ–้€ ๆˆ็š„ใ€‚ๆˆ‘ไปฌ่ฟ˜้€ๆก็œ‹ไบ†ๆœฌๆจกๅž‹ๅ’Œ GGUF ็ป“ๆžœไธๅŒ็š„ 100 ๆฎต๏ผšๅ‡ ไนŽ้ƒฝๅชๆ˜ฏๆŽช่พžไธๅŒ๏ผŒไธค่พนๅ„ๆœ‰ๅฐ‘ๆ•ฐๅฏนๆ–นๆฒกๆœ‰็š„้”™่ฏฏใ€‚

้€Ÿๅบฆ๏ผˆMacBook Pro๏ผŒM3 Max๏ผŒ128 GB๏ผŒmacOS 27.2๏ผ›้€ๆฎต็ฟป่ฏ‘๏ผ›ๆœฌๆจกๅž‹็”จ Python ็š„ mlx-lm ่ฟ่กŒ๏ผŒGGUF ็”จ llama-server ่ฟ่กŒ๏ผ‰๏ผš

GGUF Q6_K ๆœฌๆจกๅž‹
374 ๆฎตๆ€ป็”จๆ—ถ 114.6 ็ง’ 94.2 ็ง’
็”Ÿๆˆ้€Ÿๅบฆ๏ผˆไธญไฝๆ•ฐ๏ผ‰ 125 token/็ง’ 169 token/็ง’
้ฆ–ไธช token ็”จๆ—ถ๏ผˆไธญไฝๆ•ฐ๏ผ‰ 42 ๆฏซ็ง’ 35 ๆฏซ็ง’
ๅŠ ่ฝฝๆ—ถ้—ด 4.8 ็ง’ 0.9 ็ง’
ๅ†…ๅญ˜ๅณฐๅ€ผ๏ผˆ่ฟ›็จ‹๏ผ‰ 1.8 GB 2.1 GB

็ฝ‘้กต้‡Œๅพˆๅคšๆ˜ฏ็Ÿญๅฅ๏ผŒๆฏๆฌก่ฏทๆฑ‚็š„ๅ›บๅฎšๅผ€้”€ๆฏ”็”Ÿๆˆ้€Ÿๅบฆๅฝฑๅ“ๆ›ดๅคง๏ผŒๆ‰€ไปฅๆ•ดๅฅ—ๅชๅฟซ 18%ใ€‚

ๅฑ€้™

  • ๅชๅœจไธ€ๅฐๆœบๅ™จไธŠๆต‹่ฟ‡๏ผšM3 Max๏ผŒmacOS 27.2ใ€‚
  • ๅช่ƒฝ็ฟป่ฏ‘ๆ–‡ๅญ—๏ผš่ฆ็ฟป่ฏ‘ๅ›พ็‰‡่ฏท็”จๅฎ˜ๆ–นๆƒ้‡ใ€‚
  • ๆต‹่ฏ•้›†ๅพˆๅฐ๏ผš12 ไธช็ฝ‘้กตๅ…ฑ 374 ๆฎต๏ผŒไปฅๅŽŸ็‰ˆๆจกๅž‹ไธบๅ‚็…ง๏ผŒๆฒกๆœ‰็”จไบบๅทฅ่ฏ‘ๆ–‡่ฏ„ๅˆ†ใ€‚
  • ไธๆ”ฏๆŒ็นไฝ“ไธญๆ–‡๏ผšๅ’Œๅฎ˜ๆ–นๆจกๅž‹ไธ€ๆ ท๏ผŒๅฎ˜ๆ–น่ฏญ่จ€ๅˆ—่กจ้‡Œๆฒกๆœ‰็นไฝ“ไธญๆ–‡ใ€‚

่ฎธๅฏ่ฏๅ’Œๆฅๆบ

  • ่ฎธๅฏ่ฏ๏ผšApache 2.0๏ผŒๅ’Œ Index-Translate-2B ็›ธๅŒใ€‚
  • Index-Translate๏ผšๆจกๅž‹็”ฑๅ“”ๅ“ฉๅ“”ๅ“ฉ IndexTeam ๅ‘ๅธƒใ€‚
  • ๆœฌไป“ๅบ“๏ผšๆˆ‘ไปฌๅšไบ† MLX ่ฝฌๆขใ€้‡ๅŒ–ๅ’Œๆต‹่ฏ•ใ€‚ๆœฌไป“ๅบ“ไธŽ IndexTeam ๆฒกๆœ‰ๅ…ณ่”ใ€‚
Downloads last month
99
Safetensors
Model size
2B params
Tensor type
U32
ยท
BF16
ยท
MLX
Hardware compatibility
Log In to add your hardware

6-bit

Inference Providers NEW
This model isn't deployed by any Inference Provider. ๐Ÿ™‹ Ask for provider support

Model tree for OpenInfinity/Index-Translate-2B-MLX-6bit

Quantized
(24)
this model