llama.cpp/examples/simple-token-healing/README.md

2.7 KiB

llama.cpp/example/simple-token-healing

This example extends simple with token healing (aka. token alignment).

usage: ./simple-token-healing MODEL_PATH [PROMPT] [TOKEN_HEALING 0|1|d1|d|r[N]]

Examples

0: Without token healing (same as running ./simple ...):

./simple-token-healing ./models/phi-2/ggml-model-q4_0.gguf "print('Hel" 0
...
main: n_len = 32, n_ctx = 2048, n_kv_req = 32

print('Helping the customer')
...

1: Roll back the last token and constrain the bytes of the next token to start with the chopped off last token [0, 2]:

./simple-token-healing ./models/phi-2/ggml-model-q4_0.gguf "print('Hel" 1
...
token_healing: prefix = 'Hel' (1 tokens)
 [ 12621] 'Hel'
 [ 15496] 'Hello'
 [ 22087] 'Help'
 [ 28254] 'Hell'
 [ 47429] 'Helper'

main: n_len = 32, n_ctx = 2048, n_kv_req = 32

print('Hello, World!')
...

d1: Roll back multiple tokens until there doesn't exist a token which can cover the prompt's suffix and do a single constrained decoding step [2]:

./simple-token-healing ./models/phi-2/ggml-model-q4_0.gguf "print('Hello, worl" d1
...
token_healing: prefix = ' worl' (2 tokens)
 [   995] ' world'
 [  8688] ' worldwide'
 [ 11621] ' worlds'
 [ 29081] ' worldview'
 [ 43249] ' worldly'

main: n_len = 32, n_ctx = 2048, n_kv_req = 32

print('Hello, world!')
...

d: Roll back multiple tokens until there doesn't exist a token which can cover the prompt's suffix but allow multiple decoding steps:

./simple-token-healing ./models/phi-2/ggml-model-q4_0.gguf "print('Hello, worl" d
...
token_healing: prefix = ' worl' (2 tokens)

main: n_len = 32, n_ctx = 2048, n_kv_req = 32

print('Hello,
token_healing: prefix = ' worl'
 [   220] ' '
 [   266] ' w'
 [   476] ' wor'
 [   995] ' world'
 [  8688] ' worldwide'
 [ 11621] ' worlds'
 [ 24486] ' wo'
 [ 29081] ' worldview'
 [ 43249] ' worldly'
 world!')
...

r[N]: Roll back N tokens and constrain the decoding to the bytes of those tokens (multiple decoding steps) [1]. The paper [1] recommends N=3:

./simple-token-healing ./models/phi-2/ggml-model-q4_0.gguf "print('Hello, worl" r3
...
token_healing: prefix = ', worl' (3 tokens)

main: n_len = 32, n_ctx = 2048, n_kv_req = 32

print('Hello
token_healing: prefix = ', worl'
 [    11] ','
,
token_healing: prefix = ' worl'
 [   220] ' '
 [   266] ' w'
 [   476] ' wor'
 [   995] ' world'
 [  8688] ' worldwide'
 [ 11621] ' worlds'
 [ 24486] ' wo'
 [ 29081] ' worldview'
 [ 43249] ' worldly'
 world!')
...

Sources