Models
Turbo
Sign in
Download
Models
Download
Sign in
falcon3
:7b
314.9K
Downloads
Updated
8 months ago
A family of efficient AI models under 10B parameters performant in science, math, and coding through innovative training techniques.
A family of efficient AI models under 10B parameters performant in science, math, and coding through innovative training techniques.
Cancel
1b
3b
7b
10b
falcon3:7b
...
/
model
3717a52b7aea · 4.6GB
Metadata
general.architecture
llama
llama
general.file_type
Q4_K_M
Q4_K_M
llama.attention.head_count
12
12
llama.attention.head_count_kv
4
4
llama.attention.key_length
256
256
llama.attention.layer_norm_rms_epsilon
1e-06
1e-06
llama.attention.value_length
256
256
llama.block_count
28
28
llama.context_length
32768
32768
llama.embedding_length
3072
3072
llama.feed_forward_length
23040
23040
llama.rope.dimension_count
256
256
llama.rope.freq_base
1.000042e+06
1.000042e+06
llama.vocab_size
131072
131072
tokenizer.ggml.add_space_prefix
false
false
tokenizer.ggml.bos_token_id
11
11
tokenizer.ggml.eos_token_id
11
11
tokenizer.ggml.merges
[N E, Ġ Ġ, Ġ t, Ġ a, > >, ...]
[N E, Ġ Ġ, Ġ t, Ġ a, > >, ...]
tokenizer.ggml.model
gpt2
gpt2
tokenizer.ggml.padding_token_id
2023
2023
tokenizer.ggml.pre
falcon3
falcon3
tokenizer.ggml.token_type
[3, 3, 3, 3, 3, ...]
[3, 3, 3, 3, 3, ...]
tokenizer.ggml.tokens
[>>TITLE<<, >>ABSTRACT<<, >>INTRODUCTION<<, >>SUMMARY<<, >>COMMENT<<, ...]
[>>TITLE<<, >>ABSTRACT<<, >>INTRODUCTION<<, >>SUMMARY<<, >>COMMENT<<, ...]
Tensor
Name
Type
Shape
token_embd.weight
Q4_K
Q4_K
[3072, 131072]
blk.0
blk.0.attn_k.weight
Q4_K
Q4_K
[3072, 1024]
blk.0.attn_norm.weight
F32
F32
[3072]
blk.0.attn_output.weight
Q4_K
Q4_K
[3072, 3072]
blk.0.attn_q.weight
Q4_K
Q4_K
[3072, 3072]
blk.0.attn_v.weight
Q6_K
Q6_K
[3072, 1024]
blk.0.ffn_down.weight
Q6_K
Q6_K
[23040, 3072]
blk.0.ffn_gate.weight
Q4_K
Q4_K
[3072, 23040]
blk.0.ffn_norm.weight
F32
F32
[3072]
blk.0.ffn_up.weight
Q4_K
Q4_K
[3072, 23040]
blk.1
blk.1.attn_k.weight
Q4_K
Q4_K
[3072, 1024]
blk.1.attn_norm.weight
F32
F32
[3072]
blk.1.attn_output.weight
Q4_K
Q4_K
[3072, 3072]
blk.1.attn_q.weight
Q4_K
Q4_K
[3072, 3072]
blk.1.attn_v.weight
Q6_K
Q6_K
[3072, 1024]
blk.1.ffn_down.weight
Q6_K
Q6_K
[23040, 3072]
blk.1.ffn_gate.weight
Q4_K
Q4_K
[3072, 23040]
blk.1.ffn_norm.weight
F32
F32
[3072]
blk.1.ffn_up.weight
Q4_K
Q4_K
[3072, 23040]
blk.2
blk.2.attn_k.weight
Q4_K
Q4_K
[3072, 1024]
blk.2.attn_norm.weight
F32
F32
[3072]
blk.2.attn_output.weight
Q4_K
Q4_K
[3072, 3072]
blk.2.attn_q.weight
Q4_K
Q4_K
[3072, 3072]
blk.2.attn_v.weight
Q6_K
Q6_K
[3072, 1024]
blk.2.ffn_down.weight
Q6_K
Q6_K
[23040, 3072]
blk.2.ffn_gate.weight
Q4_K
Q4_K
[3072, 23040]
blk.2.ffn_norm.weight
F32
F32
[3072]
blk.2.ffn_up.weight
Q4_K
Q4_K
[3072, 23040]
blk.3
blk.3.attn_k.weight
Q4_K
Q4_K
[3072, 1024]
blk.3.attn_norm.weight
F32
F32
[3072]
blk.3.attn_output.weight
Q4_K
Q4_K
[3072, 3072]
blk.3.attn_q.weight
Q4_K
Q4_K
[3072, 3072]
blk.3.attn_v.weight
Q4_K
Q4_K
[3072, 1024]
blk.3.ffn_down.weight
Q4_K
Q4_K
[23040, 3072]
blk.3.ffn_gate.weight
Q4_K
Q4_K
[3072, 23040]
blk.3.ffn_norm.weight
F32
F32
[3072]
blk.3.ffn_up.weight
Q4_K
Q4_K
[3072, 23040]
blk.4
blk.4.attn_k.weight
Q4_K
Q4_K
[3072, 1024]
blk.4.attn_norm.weight
F32
F32
[3072]
blk.4.attn_output.weight
Q4_K
Q4_K
[3072, 3072]
blk.4.attn_q.weight
Q4_K
Q4_K
[3072, 3072]
blk.4.attn_v.weight
Q4_K
Q4_K
[3072, 1024]
blk.4.ffn_down.weight
Q4_K
Q4_K
[23040, 3072]
blk.4.ffn_gate.weight
Q4_K
Q4_K
[3072, 23040]
blk.4.ffn_norm.weight
F32
F32
[3072]
blk.4.ffn_up.weight
Q4_K
Q4_K
[3072, 23040]
blk.5
blk.5.attn_k.weight
Q4_K
Q4_K
[3072, 1024]
blk.5.attn_norm.weight
F32
F32
[3072]
blk.5.attn_output.weight
Q4_K
Q4_K
[3072, 3072]
blk.5.attn_q.weight
Q4_K
Q4_K
[3072, 3072]
blk.5.attn_v.weight
Q6_K
Q6_K
[3072, 1024]
blk.5.ffn_down.weight
Q6_K
Q6_K
[23040, 3072]
blk.5.ffn_gate.weight
Q4_K
Q4_K
[3072, 23040]
blk.5.ffn_norm.weight
F32
F32
[3072]
blk.5.ffn_up.weight
Q4_K
Q4_K
[3072, 23040]
blk.6
blk.6.attn_k.weight
Q4_K
Q4_K
[3072, 1024]
blk.6.attn_norm.weight
F32
F32
[3072]
blk.6.attn_output.weight
Q4_K
Q4_K
[3072, 3072]
blk.6.attn_q.weight
Q4_K
Q4_K
[3072, 3072]
blk.6.attn_v.weight
Q4_K
Q4_K
[3072, 1024]
blk.6.ffn_down.weight
Q4_K
Q4_K
[23040, 3072]
blk.6.ffn_gate.weight
Q4_K
Q4_K
[3072, 23040]
blk.6.ffn_norm.weight
F32
F32
[3072]
blk.6.ffn_up.weight
Q4_K
Q4_K
[3072, 23040]
blk.7
blk.7.attn_k.weight
Q4_K
Q4_K
[3072, 1024]
blk.7.attn_norm.weight
F32
F32
[3072]
blk.7.attn_output.weight
Q4_K
Q4_K
[3072, 3072]
blk.7.attn_q.weight
Q4_K
Q4_K
[3072, 3072]
blk.7.attn_v.weight
Q4_K
Q4_K
[3072, 1024]
blk.7.ffn_down.weight
Q4_K
Q4_K
[23040, 3072]
blk.7.ffn_gate.weight
Q4_K
Q4_K
[3072, 23040]
blk.7.ffn_norm.weight
F32
F32
[3072]
blk.7.ffn_up.weight
Q4_K
Q4_K
[3072, 23040]
blk.8
blk.8.attn_k.weight
Q4_K
Q4_K
[3072, 1024]
blk.8.attn_norm.weight
F32
F32
[3072]
blk.8.attn_output.weight
Q4_K
Q4_K
[3072, 3072]
blk.8.attn_q.weight
Q4_K
Q4_K
[3072, 3072]
blk.8.attn_v.weight
Q6_K
Q6_K
[3072, 1024]
blk.8.ffn_down.weight
Q6_K
Q6_K
[23040, 3072]
blk.8.ffn_gate.weight
Q4_K
Q4_K
[3072, 23040]
blk.8.ffn_norm.weight
F32
F32
[3072]
blk.8.ffn_up.weight
Q4_K
Q4_K
[3072, 23040]
blk.9
blk.9.attn_k.weight
Q4_K
Q4_K
[3072, 1024]
blk.9.attn_norm.weight
F32
F32
[3072]
blk.9.attn_output.weight
Q4_K
Q4_K
[3072, 3072]
blk.9.attn_q.weight
Q4_K
Q4_K
[3072, 3072]
blk.9.attn_v.weight
Q4_K
Q4_K
[3072, 1024]
blk.9.ffn_down.weight
Q4_K
Q4_K
[23040, 3072]
blk.9.ffn_gate.weight
Q4_K
Q4_K
[3072, 23040]
blk.9.ffn_norm.weight
F32
F32
[3072]
blk.9.ffn_up.weight
Q4_K
Q4_K
[3072, 23040]
blk.10
blk.10.attn_k.weight
Q4_K
Q4_K
[3072, 1024]
blk.10.attn_norm.weight
F32
F32
[3072]
blk.10.attn_output.weight
Q4_K
Q4_K
[3072, 3072]
blk.10.attn_q.weight
Q4_K
Q4_K
[3072, 3072]
blk.10.attn_v.weight
Q4_K
Q4_K
[3072, 1024]
blk.10.ffn_down.weight
Q6_K
Q6_K
[23040, 3072]
blk.10.ffn_gate.weight
Q4_K
Q4_K
[3072, 23040]
blk.10.ffn_norm.weight
F32
F32
[3072]
blk.10.ffn_up.weight
Q4_K
Q4_K
[3072, 23040]
blk.11
blk.11.attn_k.weight
Q4_K
Q4_K
[3072, 1024]
blk.11.attn_norm.weight
F32
F32
[3072]
blk.11.attn_output.weight
Q4_K
Q4_K
[3072, 3072]
blk.11.attn_q.weight
Q4_K
Q4_K
[3072, 3072]
blk.11.attn_v.weight
Q4_K
Q4_K
[3072, 1024]
blk.11.ffn_down.weight
Q4_K
Q4_K
[23040, 3072]
blk.11.ffn_gate.weight
Q4_K
Q4_K
[3072, 23040]
blk.11.ffn_norm.weight
F32
F32
[3072]
blk.11.ffn_up.weight
Q4_K
Q4_K
[3072, 23040]
blk.12
blk.12.attn_k.weight
Q4_K
Q4_K
[3072, 1024]
blk.12.attn_norm.weight
F32
F32
[3072]
blk.12.attn_output.weight
Q4_K
Q4_K
[3072, 3072]
blk.12.attn_q.weight
Q4_K
Q4_K
[3072, 3072]
blk.12.attn_v.weight
Q6_K
Q6_K
[3072, 1024]
blk.12.ffn_down.weight
Q4_K
Q4_K
[23040, 3072]
blk.12.ffn_gate.weight
Q4_K
Q4_K
[3072, 23040]
blk.12.ffn_norm.weight
F32
F32
[3072]
blk.12.ffn_up.weight
Q4_K
Q4_K
[3072, 23040]
blk.13
blk.13.attn_k.weight
Q4_K
Q4_K
[3072, 1024]
blk.13.attn_norm.weight
F32
F32
[3072]
blk.13.attn_output.weight
Q4_K
Q4_K
[3072, 3072]
blk.13.attn_q.weight
Q4_K
Q4_K
[3072, 3072]
blk.13.attn_v.weight
Q4_K
Q4_K
[3072, 1024]
blk.13.ffn_down.weight
Q6_K
Q6_K
[23040, 3072]
blk.13.ffn_gate.weight
Q4_K
Q4_K
[3072, 23040]
blk.13.ffn_norm.weight
F32
F32
[3072]
blk.13.ffn_up.weight
Q4_K
Q4_K
[3072, 23040]
blk.14
blk.14.attn_k.weight
Q4_K
Q4_K
[3072, 1024]
blk.14.attn_norm.weight
F32
F32
[3072]
blk.14.attn_output.weight
Q4_K
Q4_K
[3072, 3072]
blk.14.attn_q.weight
Q4_K
Q4_K
[3072, 3072]
blk.14.attn_v.weight
Q4_K
Q4_K
[3072, 1024]
blk.14.ffn_down.weight
Q4_K
Q4_K
[23040, 3072]
blk.14.ffn_gate.weight
Q4_K
Q4_K
[3072, 23040]
blk.14.ffn_norm.weight
F32
F32
[3072]
blk.14.ffn_up.weight
Q4_K
Q4_K
[3072, 23040]
blk.15
blk.15.attn_k.weight
Q4_K
Q4_K
[3072, 1024]
blk.15.attn_norm.weight
F32
F32
[3072]
blk.15.attn_output.weight
Q4_K
Q4_K
[3072, 3072]
blk.15.attn_q.weight
Q4_K
Q4_K
[3072, 3072]
blk.15.attn_v.weight
Q6_K
Q6_K
[3072, 1024]
blk.15.ffn_down.weight
Q4_K
Q4_K
[23040, 3072]
blk.15.ffn_gate.weight
Q4_K
Q4_K
[3072, 23040]
blk.15.ffn_norm.weight
F32
F32
[3072]
blk.15.ffn_up.weight
Q4_K
Q4_K
[3072, 23040]
blk.16
blk.16.attn_k.weight
Q4_K
Q4_K
[3072, 1024]
blk.16.attn_norm.weight
F32
F32
[3072]
blk.16.attn_output.weight
Q4_K
Q4_K
[3072, 3072]
blk.16.attn_q.weight
Q4_K
Q4_K
[3072, 3072]
blk.16.attn_v.weight
Q4_K
Q4_K
[3072, 1024]
blk.16.ffn_down.weight
Q6_K
Q6_K
[23040, 3072]
blk.16.ffn_gate.weight
Q4_K
Q4_K
[3072, 23040]
blk.16.ffn_norm.weight
F32
F32
[3072]
blk.16.ffn_up.weight
Q4_K
Q4_K
[3072, 23040]
blk.17
blk.17.attn_k.weight
Q4_K
Q4_K
[3072, 1024]
blk.17.attn_norm.weight
F32
F32
[3072]
blk.17.attn_output.weight
Q4_K
Q4_K
[3072, 3072]
blk.17.attn_q.weight
Q4_K
Q4_K
[3072, 3072]
blk.17.attn_v.weight
Q4_K
Q4_K
[3072, 1024]
blk.17.ffn_down.weight
Q4_K
Q4_K
[23040, 3072]
blk.17.ffn_gate.weight
Q4_K
Q4_K
[3072, 23040]
blk.17.ffn_norm.weight
F32
F32
[3072]
blk.17.ffn_up.weight
Q4_K
Q4_K
[3072, 23040]
blk.18
blk.18.attn_k.weight
Q4_K
Q4_K
[3072, 1024]
blk.18.attn_norm.weight
F32
F32
[3072]
blk.18.attn_output.weight
Q4_K
Q4_K
[3072, 3072]
blk.18.attn_q.weight
Q4_K
Q4_K
[3072, 3072]
blk.18.attn_v.weight
Q6_K
Q6_K
[3072, 1024]
blk.18.ffn_down.weight
Q4_K
Q4_K
[23040, 3072]
blk.18.ffn_gate.weight
Q4_K
Q4_K
[3072, 23040]
blk.18.ffn_norm.weight
F32
F32
[3072]
blk.18.ffn_up.weight
Q4_K
Q4_K
[3072, 23040]
blk.19
blk.19.attn_k.weight
Q4_K
Q4_K
[3072, 1024]
blk.19.attn_norm.weight
F32
F32
[3072]
blk.19.attn_output.weight
Q4_K
Q4_K
[3072, 3072]
blk.19.attn_q.weight
Q4_K
Q4_K
[3072, 3072]
blk.19.attn_v.weight
Q4_K
Q4_K
[3072, 1024]
blk.19.ffn_down.weight
Q4_K
Q4_K
[23040, 3072]
blk.19.ffn_gate.weight
Q4_K
Q4_K
[3072, 23040]
blk.19.ffn_norm.weight
F32
F32
[3072]
blk.19.ffn_up.weight
Q4_K
Q4_K
[3072, 23040]
blk.20
blk.20.attn_k.weight
Q4_K
Q4_K
[3072, 1024]
blk.20.attn_norm.weight
F32
F32
[3072]
blk.20.attn_output.weight
Q4_K
Q4_K
[3072, 3072]
blk.20.attn_q.weight
Q4_K
Q4_K
[3072, 3072]
blk.20.attn_v.weight
Q6_K
Q6_K
[3072, 1024]
blk.20.ffn_down.weight
Q6_K
Q6_K
[23040, 3072]
blk.20.ffn_gate.weight
Q4_K
Q4_K
[3072, 23040]
blk.20.ffn_norm.weight
F32
F32
[3072]
blk.20.ffn_up.weight
Q4_K
Q4_K
[3072, 23040]
blk.21
blk.21.attn_k.weight
Q4_K
Q4_K
[3072, 1024]
blk.21.attn_norm.weight
F32
F32
[3072]
blk.21.attn_output.weight
Q4_K
Q4_K
[3072, 3072]
blk.21.attn_q.weight
Q4_K
Q4_K
[3072, 3072]
blk.21.attn_v.weight
Q4_K
Q4_K
[3072, 1024]
blk.21.ffn_down.weight
Q4_K
Q4_K
[23040, 3072]
blk.21.ffn_gate.weight
Q4_K
Q4_K
[3072, 23040]
blk.21.ffn_norm.weight
F32
F32
[3072]
blk.21.ffn_up.weight
Q4_K
Q4_K
[3072, 23040]
blk.22
blk.22.attn_k.weight
Q4_K
Q4_K
[3072, 1024]
blk.22.attn_norm.weight
F32
F32
[3072]
blk.22.attn_output.weight
Q4_K
Q4_K
[3072, 3072]
blk.22.attn_q.weight
Q4_K
Q4_K
[3072, 3072]
blk.22.attn_v.weight
Q4_K
Q4_K
[3072, 1024]
blk.22.ffn_down.weight
Q4_K
Q4_K
[23040, 3072]
blk.22.ffn_gate.weight
Q4_K
Q4_K
[3072, 23040]
blk.22.ffn_norm.weight
F32
F32
[3072]
blk.22.ffn_up.weight
Q4_K
Q4_K
[3072, 23040]
blk.23
blk.23.attn_k.weight
Q4_K
Q4_K
[3072, 1024]
blk.23.attn_norm.weight
F32
F32
[3072]
blk.23.attn_output.weight
Q4_K
Q4_K
[3072, 3072]
blk.23.attn_q.weight
Q4_K
Q4_K
[3072, 3072]
blk.23.attn_v.weight
Q6_K
Q6_K
[3072, 1024]
blk.23.ffn_down.weight
Q6_K
Q6_K
[23040, 3072]
blk.23.ffn_gate.weight
Q4_K
Q4_K
[3072, 23040]
blk.23.ffn_norm.weight
F32
F32
[3072]
blk.23.ffn_up.weight
Q4_K
Q4_K
[3072, 23040]
blk.24
blk.24.attn_k.weight
Q4_K
Q4_K
[3072, 1024]
blk.24.attn_norm.weight
F32
F32
[3072]
blk.24.attn_output.weight
Q4_K
Q4_K
[3072, 3072]
blk.24.attn_q.weight
Q4_K
Q4_K
[3072, 3072]
blk.24.attn_v.weight
Q6_K
Q6_K
[3072, 1024]
blk.24.ffn_down.weight
Q6_K
Q6_K
[23040, 3072]
blk.24.ffn_gate.weight
Q4_K
Q4_K
[3072, 23040]
blk.24.ffn_norm.weight
F32
F32
[3072]
blk.24.ffn_up.weight
Q4_K
Q4_K
[3072, 23040]
blk.25
blk.25.attn_k.weight
Q4_K
Q4_K
[3072, 1024]
blk.25.attn_norm.weight
F32
F32
[3072]
blk.25.attn_output.weight
Q4_K
Q4_K
[3072, 3072]
blk.25.attn_q.weight
Q4_K
Q4_K
[3072, 3072]
blk.25.attn_v.weight
Q6_K
Q6_K
[3072, 1024]
blk.25.ffn_down.weight
Q6_K
Q6_K
[23040, 3072]
blk.25.ffn_gate.weight
Q4_K
Q4_K
[3072, 23040]
blk.25.ffn_norm.weight
F32
F32
[3072]
blk.25.ffn_up.weight
Q4_K
Q4_K
[3072, 23040]
blk.26
blk.26.attn_k.weight
Q4_K
Q4_K
[3072, 1024]
blk.26.attn_norm.weight
F32
F32
[3072]
blk.26.attn_output.weight
Q4_K
Q4_K
[3072, 3072]
blk.26.attn_q.weight
Q4_K
Q4_K
[3072, 3072]
blk.26.attn_v.weight
Q6_K
Q6_K
[3072, 1024]
blk.26.ffn_down.weight
Q6_K
Q6_K
[23040, 3072]
blk.26.ffn_gate.weight
Q4_K
Q4_K
[3072, 23040]
blk.26.ffn_norm.weight
F32
F32
[3072]
blk.26.ffn_up.weight
Q4_K
Q4_K
[3072, 23040]
blk.27
blk.27.attn_k.weight
Q4_K
Q4_K
[3072, 1024]
blk.27.attn_norm.weight
F32
F32
[3072]
blk.27.attn_output.weight
Q4_K
Q4_K
[3072, 3072]
blk.27.attn_q.weight
Q4_K
Q4_K
[3072, 3072]
blk.27.attn_v.weight
Q6_K
Q6_K
[3072, 1024]
blk.27.ffn_down.weight
Q6_K
Q6_K
[23040, 3072]
blk.27.ffn_gate.weight
Q4_K
Q4_K
[3072, 23040]
blk.27.ffn_norm.weight
F32
F32
[3072]
blk.27.ffn_up.weight
Q4_K
Q4_K
[3072, 23040]
output.weight
Q6_K
Q6_K
[3072, 131072]
output_norm.weight
F32
F32
[3072]