Modelfile-llama3.2-vision

#  ollama create llama-3.2-90b-instruct:q4_K_M -f modelfiles/Modelfile-llama3.2-vision

#  FROM ../Llama-3.2-11B-Instruct-Q8_0.gguf
FROM llama3.2-vision:90b-instruct-q4_K_M

#  template version 715415638c9c 2024-11-20
TEMPLATE """{{- range $index, $_ := .Messages }}<|start_header_id|>{{ .Role }}<|end_header_id|>

{{ .Content }}
{{- if gt (len (slice $.Messages $index)) 1 }}<|eot_id|>
{{- else if ne .Role "assistant" }}<|eot_id|><|start_header_id|>assistant<|end_header_id|>

{{ end }}
{{- end }}"""

PARAMETER stop "<|start_header_id|>"
PARAMETER stop "<|end_header_id|>"
PARAMETER stop "<|eot_id|>"

PARAMETER num_ctx 32768
PARAMETER num_batch 1024


### min_p sampling ##
# min_p works best with a bit of temperature
PARAMETER temperature 0.6
# 1.0 disables top_p, so we can use min_p
PARAMETER top_p 1.0
PARAMETER min_p 0.9
### min_p sampling ##

#  PARAMETER temperature 0.6
#  PARAMETER top_p 0.9